Particle induced x-ray emission study of rod outer segment disc membranes


Material Information

Particle induced x-ray emission study of rod outer segment disc membranes
Physical Description:
viii, 85 leaves : ill. ; 28 cm.
McCormick, Larry Don, 1944-
Publication Date:


Subjects / Keywords:
Visual pigments   ( lcsh )
Retina   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Includes bibliographical references (leaves 80-84).
Statement of Responsibility:
by Larry Don McCormick.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000092369
notis - AAK7775
oclc - 05955522
System ID:

This item is only available as the following downloads:

Full Text








I would like to express my gratitude to the many persons

who assisted me at various times. In particular, the time and

effort of my committee members are greatly appreciated. In

addition, the members of the accelerator group furnished in-

valuable help. Numerous persons assisted in making this a

truly interdisciplinary and even interuniversity and interstate

effort. Foremost among these are Drs. Robert Cohen and Phillip

Achey who allowed the use of their laboratory facilities and

Henry Kaufman who furnished the necessary data fitting program.

Finally, I want to especially thank my wife for her extreme

patience and understanding during the last few years.


ACKNOWLEDGEMENTS .......................

LIST OF TABLES .........................

LIST OF FIGURES........................

ABSTRACT ...............................


I INTRODUCTION..................



















Jt ., 'C_7l\ UJLi j .LiJ, L LJ L I- JtiL- L LJ L- i iUJ-.L L ... .. .. .




EXPERIMENTAL RESULTS ...........................


SAMPLES ........................................


MEMBRANES ....................................


H E X . . . . . .

MITOCHONDRIA ASSAY ...................... ......






BIBLIOGRAPHY ......................................... 80

BIOGRAPHICAL SKETCH ..................................... 85




IN PIXE....................................... 29

IN PIXE.......................................... 32

TARGETS ....................................... 59

DISC MEMBRANES................................ 60




APPROXIMATION................................. 25




6 X-RAY SPECTRUM OF FICOLL ...................... 53

7 X-RAY SPECTRUM OF SUCROSE ..................... 54

BUFFER......................................... 55

EXTERNAL ABSORBER ............................. 56

EXTERNAL ABSORBER ............................. 57

Abstract of Dissertation Presented to the
Graduate Council of the University of Florida
in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy



Larry Don McCormick

June, 1979

Chairman: Henri A. Van Rinsvelt
Cochairman: F. Eugene Dunnam
Major Department: Physics and Astronomy

A relatively new physics technique, particle induced X-ray

emission (PIXE), was used to determine the trace element con-

tent of bovine retinal disc membranes. In this technique,

characteristic X-rays resulting from inner shell atomic vacancy

production by heavy, charged particles are detected. With

standardization the amount of each element in a sample can

be determined. PIXE is multi-elemental allowing the simulta-

neous determination of all trace elements and making it a very

useful biological technique. The multi-elemental property of

PIXE was used in this study to determine the trace element

content per rhodopsin without performing a separate rhodopsin

assay. First, the ratio of each element to sulfur in a sample

was determined. This ratio was then related to the amount of

rhodopsin using a sulfur content of ten per rhodopsin. Disc

membranes were obtained by a standard separation procedure.


The amount of trace elements per rhodopsin were as

follows: calcium .492 .153, iron .0243 .0075, copper

.132 .060, and zinc .0577 .010. In addition chlorine was

present at a level of .903 .445 per rhodopsin. Commonly

occurring biological metals which were not detected are man-

ganese, molybdenum and nickel. Upper limits determined per

rhodopsin for these elements are .007, .0060 and .0033 re-

spectively. No concentration changes of any of the detected

elements were noted after light exposure of disc membrane

samples. Bleaching levels used were between .1% and 100%.




Sensory receptors are highly specialized cells which cap-

ture energy and initiate an impulse in adjoining nerve cells.

This process is referred to as transduction. For each type

of receptor cell, researchers are attempting to isolate and

subsequently study the molecule responsible for energy absorp-

tion. Concurrent with this they are trying to understand the

transduction process.

In the visual process, the transducing molecule has been

identified and extensively studied.1-4 Rhodopsin is the mole-

cule responsible for absorbing light and beginning the visual

chain of events. Vision is unique among the senses, because

no other transducing molecules have been discovered.

In 1876, Franz Boll successfully demonstrated the presence

of a visual pigment in a freshly isolated retina, and noted that

light incident upon the retina caused the characteristic red

color to change to white.5 This process is referred to as


The role of rhodopsin in transduction was the subject of

many brilliant studies conducted by Wald and co-workers.6-7

It is largely due to these studies that visual transduction

has been characterized on a molecular level. As a result, it

has been widely studied by biophysicists attempting to discover

the mechanism for the conversion of light energy into a nerve

impulse. This dissertation is the report of one such study

of the visual transduction mechanism.



The posterior portion of vertebrate eyes, the retina,

often contain two different types of visual receptor cells,

rods and cones. Each cell is named because of its distinc-

tive shape. Cones are responsible for color vision, whereas,

rods are primarily useful at low light intensities. Rods con-

sist of two distinct sections, the inner and outer segment.

The inner segment contains organelles which maintain the cell.

In contrast, the outer segment consists almost entirely of flattened

sacs formed by two parallel disc shaped membranes. Most of

the cell's rhodopsin is contained in the disc membranes. Thus,

it is expected that this portion of the cell is responsible

for transduction.

Rod outer segments are large on a cellular scale. Bo-

vine rod outer segments are approximately 7-10 micrometers in

length and 2 micrometers in diameter8 resulting in a volume

as large as 3 x 1014 liters. Their large size enables cell

physiologists to pierce the external (or plasma) membrane and

monitor the transmembrane electrical signals. This type of
2 3,5
research has yielded many results.' Even though these studies

cannot determine the mechanism of transduction, any theory

attempting to explain transduction must agree with the results


Using thin section electron microscopy, Mason has

determined a typical disc to disc repeat distance of 250 ang-

stroms and an intradisc space of 25 angstroms. Each disc mem-

brane has a thickness of 75-100 angstroms. The membrane forms

a sac and each bovine rod outer segment contains approximately

450 discs.9 The dimensions of discs in other vertebrates as

determined by electron microscopy and X-ray diffraction are

similar.10 (See Figure 1.)

The sacs are layered or stacked in the rod outer segment

and surrounded by the plasma membrane. The disc plane is per-

pendicular to the rod's longitudinal axis. Electron micro-

graphs indicate there is no contact between the disc and plasma
membrane. It should be noted that electron microscopy re-

quires extensive sample preparation and the results are not

always considered conclusive.

The biogenesis of disc membranes is unclear. Mason has

indicated that the discs are formed as the result of fusion of

small vesicles observed at the base of the rod outer segment

near the cilium connecting outer and inner segments. However,

Young and other researchers have noted what he described as

invaginations of the outer segment plasma membrane, forming

layered discs.11

It is generally agreed that rhodopsin is synthesized in

the ribosomes of the inner segment, afterward migrating to the

Golgi apparatus. However, the role of the Golgi apparatus in

this sequence is unknown.12 Subsequently, rhodopsin migrates

through the connecting cilium and is placed in the discs. The

details of rhodopsin synthesis and incorporation into the disc


membranes are currently the object of much research by bio-

chemists and cell physiologists.

After formation, the discs move toward the posterior of

the outer segment, eventually being phagocytized by the pig-

ment epithelium. The process typically requires several weeks

but varies considerably among species and with the amount of

light exposure.

The study of the planar arrangement of discs has been

accomplished using low angle X-ray diffraction and neutron

scattering. Each technique indicates that rhodopsin is lo-

cated in each membrane of the disc bilayer and has a rela-

tively uniform cross section which completely spans a membrane.13

Raman and circular dichroism studies of disc membranes

suggest that most of the protein is alpha-helical, indicating

a high degree of organization. 1415 This is interpretated

by some researchers to imply the molecule forms a "channel"

spanning the disc membrane. This is of importance in the

currently accepted models of the transduction mechanism.

The biochemical composition of disc membranes is well

known and unusual; because, the protein component is domi-

nated by one protein, rhodopsin, which is 87% of the disc

protein. The remaining protein components are not as well

known. Lipids comprise 50% of the disc membranes. Protein

accounts for 36;', and carbohydrates make up 4%.16

Rhodopsin is a combination of the membrane protein opsin

and the chromophore 11-cis retinal. Retinal is probably

bound to opsin with an e-nitrogen of lysine.17 One important

unexplained problem of vision is how this bonding yields a


maximum absorption above 500 nanometers instead of 440 nano-

meters which would be expected based upon similar compounds.



The excitation process in a vertebrate rod is known to

begin with a conformational change in rhodopsin, a photochem-

ical cis-trans isomerization of retinal, and conclude in a

transient change in the plasma membrane potential.18

The mechanism linking these two well studied processes

has been the subject of many studies and speculations. Since

this linking mechanism is the object of this study, some back-

ground experimental results and the currently accepted theory

will be briefly presented.

One of the most striking aspects of vision is the small

"amount" of light required. The threshold for visual excitation

may be as low as one photon7 and only 1% bleaching is required

for saturation response. This results in an amplification of

about one million. Since only one rhodopsin is required to ab-

sorb a photon, each rhodopsin must somehow be linked to the output.

As previously mentioned, there appears to be no connection

between the disc membrane containing rhodopsin and the plasma

membrane. Thus, signals must be transmitted across the cell's

cytoplasmic space. The distance from an activated rhodopsin

to the plasma membrane can be as large as one micrometer.

Vertebrate rods containing large amounts of rhodopsin,

typically 10 molecules.19 This indicates the excitation mech-

anism is capable of detecting much less than one part per million

in altered receptors.

The result of transduction is a hyperpolarization (an in-

crease in the negative potential across the plasma membrane) instead of

depolarization. This is in contrast to the depolarization of

nerve cells which is much better understood. A theory is re-

quired which somehow explains an increase in trans membrane


The currently accepted theory of the ionic mechanism of
transduction is due to Hagins and Yoshikami.9 They formulated

their theory after having performed an extensive series of

experiments dealing with the effect of various ions upon the

plasma membrane potential. Originally, they had noted that

when a rod is dark-adapted, current passively flows through the

plasma membrane into the rod outer segment. The current is

extruded somewhere along the inner segment. The inward flow

results from the potential difference of approximately 50 milli-

volts (outside positive) across the plasma membrane. Researchers

have not been able to further localize the sites of current

flow. By varying the rod's bathing medium, Hagins and Yoshi-

kami were able to show that Na+ was required externally or the

rod's response to light was abolished. However, upon allowing

Na+ to be replaced in the bathing medium, the response to

light was completely renewed. Based upon this result, they

concluded that the dark current is due to Na+ flow, and the

flow is somehow suppressed as the result of light absorption

in the discs. Since the efflux of ia+ continues in the inner

segment, drawing Na+ from the outer segment, the net result is

an increase in membrane potential. (See Figure 1)




w 0

r- \

/- Z


0(- O


0 CO

In an attempt to determine how the dark current into the

outer segment is decreased, Hagins and Yoshikami injected

various ions into the rod's cytoplasm. The result was that

the addition of Ca+ had the same effect as light absorption,

but the intracellular addition of K Cl Na or Mg had

no effect upon the membrane potential consistent with the ab-

sorption of light.

Due to the marked effects of Ca+ on both the dark current

and trans membrane potential, Hagins and Yoshikami proposed

that the dark current enters the outer segment through Na+
specific channels located near possible Ca4 binding sites on

the cytoplasmic (inner) side of the plasma membrane. If a Ca+

can prevent the influx of many Na+ ions, the large amplifica-

tion is easily explained.20

To complete their theory, Hagins and Yoshikami proposed

that Ca+ is initially sequestered inside the discs. The absorp-

tion of a photon by rhodopsin releases a number (3-300) of Ca+

ions from the discs into the rod's cytoplasm. Subsequently

Ca++ diffuses to the plasma membrane, attaches at the appropri-

ate sites, and reduces the Na+ influx. Since the cytoplasmic

side of the plasma membrane is already negative relative to the

exterior, a reduction in Na influx results in an increased

trans membrane potential, a hyperpolarization.



Since Hagins and Yoshikami proposed their ionic theory

of hyperpolarization, there have been numerous attempts to

experimentally prove or disprove this theory. These experi-

ments and others which indirectly yield evidence affecting

their theory, will be discussed in this chapter.

Smith et al. examined the light-activated release of Ca+
from sonicated discs.1 They separated intact discs from bo-

vine retinas, then placed the discs in a solution containing

radioactive Ca-. The solution was subjected to ultrasonic

waves (sonicated), which tends to make the membranes leak.

During this time, the radioactive Ca+ should have diffused

into the discs. After sonication, the discs are stored allow-

ing membrane resealing and retaining some of the radioactive

Ca The researchers then subjected the resealed discs to

light while the effluent was being collected. Rhodopsin

bleaching levels were between 24% and 57% and indicated a

release of approximately 1 Ca+ per rhodopsin bleached. If

the discs were not sonicated, they could not obtain any ex-

change with radioactive Ca Therefore, they concluded that

sonication is required to load discs with Ca-. However, the

authors were not able to show any Ca+ efflux with low

light levels nor were they able to demonstrate that the discs

were not disrupted by sonication.

Mason et al. performed a similar experiment.22 They ob-

tained purified rod outer segments which were sonicated, incu-

bated in a radioactive Ca solution, and resealed as vesicles.

The authors claimed the resealed vesicles were similar to discs

in appearance in electron micrographs. They were able to mea-

sure the ratio of Ca released to rhodopsin bleached with

light levels between 5,000 photons per rod and 100% bleaching

to be close to 1. In addition, they were able to show that

sonicated discs actively accumulated Ca against a concentra-

tion gradient. This accumulation would be required in order

to replenish the intradisc Ca+ supply.

The most extensive study utilizing radioactive tracers

was performed by Sorbi and Cavaggioni.23 They obtained frog

rod outer segments which were passed through a microsyringe

needle in order to burst the plasma membrane. The suspension

was then placed in a medium containing radioactive ions.

They used radioactive K Na, Rb, Cl and Ca and were

able to load the tracers without sonication. After loading,

the discs were washed and subjected to varying amounts of lumi-

nous flux, ranging from a few photons per second per disc to

almost saturation levels. Aliquots of the effluent were col-

lected and analyzed for radioactivity. Low light levels did

not alter the efflux of any of the radioactive ions. High

light levels (approximately 50% bleaching) increased the efflux

of Na Rb and Cl. The authors detected no change in Ca+

loss due to bleaching at any light level. Their experimental

results placed an upper limit for Ca+ release due to light at

a few hundred Ca+ ions released per disc. Thus, at high

bleaching levels they should have been able to detect the Ca+4

changes predicted by Hagins and Yoshikami and a release of 1

Ca++ per rhodopsin bleached.

Liebman24 used atomic absorption spectroscopy in an at-

tempt to show a direct effect of light upon the Ca+ content

of disc membranes. He obtained rod outer segments which were

fragmented by vortexing in a medium containing Ca This left

only the disc compartment intact. Subsequently, the disc sus-

pension was completely bleached or kept in the dark. He then

determined Ca+ content by atomic absorption spectroscopy.

His results indicated a Ca++ content which increased with in-

creasing Ca+ concentration in the incubation medium and de-

creased with light exposure. For bovine rod outer segments

in a 2 micromolar Ca solution, there were approximately 3.3

Ca+ per rhodopsin after bleaching and 4 Ca per rhodopsin

before bleaching. However, the author felt that the time de-

lay between bleaching and target preparation allowed re-entry

of Ca+ into the discs. Subsequently he repeated the experi-

ment with frog outer segments except that the medium contained

a Ca chelating agent (an anion which strongly binds cations)

which would prevent its re-entry. The results were that .2

Ca+ were released per rhodopsin bleached. It should be noted

that these results were obtained with light levels much higher

than physiological and yield a Ca release at least an order

of magnitude lower than required by the theory.

An extensive study which was designed to use low light

levels was completed by Szuts and Cone in 1977.25 They used

dark-adapted bullfrog retinas, measuring the Ca4+ and Mg++

contents of both discs and rod outer segments and attempting

to detect changes due to physiological light levels. The

light levels used ranged from approximately .001% to 1% bleach-

ing. The authors designed their experiment to be done at low

levels because rods function at low levels and are overridden

by cone functions at higher light levels.

Retinas were shaken into a sucrose medium yielding outer

segments which were exposed to the appropriate amount of light.

(In some instances, Ca was added to the sucrose medium.)

Immediately after exposure, a hypotonic solution containing a

chelating anion was added to burst the outer segment plasma

membrane and chelate any released Ca The suspension was

centrifuged to pellet the discs and the pellet prepared for

atomic absorption spectroscopy measurements. Results indicated

that neither disc nor rod Ca levels depended upon Ca+ levels

in the external medium. This is in marked disagreement with

Liebman's results.24 Calcium content of discs was found to

be approximately 1 Ca+ per rhodopsin. They were not able

to detect any statistically significant change in Ca+ or Mg+

due to light absorption. They even found that in some samples

Ca+ levels were higher in samples exposed to light than in

samples which remained in the dark. The authors felt that

they could have detected a 20% change in Ca content. Since

this is much higher than that required to satisfy Hagins and

Yoshikami's theory, they concluded that they were not able to

definitively test the theory.

A similar experiment conducted by Hendriks et al.26 was

done at higher light levels and indicated Ca+ release upon

light exposure. These researchers used bleaching levels great-

er than 50%. They found that Ca++ content in dark-adapted frog

rod outer segments was 11 Ca+ per rhodopsin if they were sep-

arated in a solution containing 3 millimolar adenosine triphos-

phate (ATP). The outer segments were either lysed and bleached

or bleached and then lysed. In both cases, a 20% loss of Ca44

was reported. For their reported Ca++ levels, this is a 3 Ca44

released per rhodopsin bleached.

An interesting experiment done by Tam et al.27 indicates

Zn may be an important trace element in the functioning of

visual excitation. Initially, the authors extracted rhodopsin

utilizing the detergent emulphogene. The extract was exposed

to light and then subjected to column chromatography which

separates out the rhodopsin fraction. This fraction was ana-

lyzed for Mg Ca and Zn content. The results were 1.4

Ca 1.4 Mg+ and .8 Zn+ per rhodopsin in non-exposed sam-

ples. The levels of Ca+ and Hg were not dependent upon

illumination; however levels of Zn++ in illuminated samples

were 60% higher than in non-illuminated samples. Bleaching

was 100% since the authors subjected the samples to light un-

til no absorption at 498 nanometers was detected in the rho-

dopsin fraction. The value of .8 Zn per rhodopsin clearly

is not stoichiometric, but the authors feel this is due to

Zn losses in handling.

That rhodopsin is not a metalloprotein had been demon-

strated many years prior by Fukami et al.28 For their study,

they used frozen retinas. Tam's group found that if they froze

their preparation at any stage, Zn++ was completely lost.

These results appear to reopen the question of metallic ions

bound to rhodopsin.

In addition to the previously described experiments

attempting to directly measure Ca++ flux from photoreceptor

membranes due to light exposure, there are numerous studies

which indirectly indicate Ca+ may be involved in transduction.

Neufeld et al.29 studied the binding of radioactive Ca++

to sonicated bovine rod outer segments. They found that Ca4+

binding varied with incubation conditions; however, no light

effect could be detected. The addition of ATP to the external

medium more than doubled the amount of Ca bound. This re-

sult would be expected if disc membranes contain a Ca+ -ATPase

for the active transport of Ca4+ back into the discs.
Hemminki utilized intact bovine rod outer segments which

had been subjected to light or kept in the dark. Subsequently,

all samples were placed in a medium containing radioactive
Ca He found that rods which had remained in the dark bound

more Ca than rods which had been exposed to light.

Brown and Flaming31 repeated that portion of Hagins and

Yoshimaki's experiment dealing with the effect of ions injected

into functional rod outer segments. They studied the effect

of injected Sr and Ba upon the rod outer segment plasma

membrane potential. Sr had no effect; however, Ba re-

versed the effect of cytoplasmic Ca An increase in intra-

cellular Ba+ concentration increased the sensitivity of rods

to light and more light was required to yield an equivalent

hyperpolarization of the plasma membrane. The authors state

that Ba+ has been detected in vertebrate retina and indicate

that it might have physiological significance. However,

the presence of Ba in vertebrate photoreceptors has not

been demonstrated.

One study which may contradict Hagins and Yoshikami was
performed by Hess.32 She exposed frogs to light in vivo and

compared Ca content of rod outer segments extracted from

dark-adapted and light-adapted frogs. In addition, she ex-

posed retinas extracted from in vivo dark-adapted frogs. She

found that rods from in vivo light-adapted retinas contained

4 Ca per rhodopsin compared to .25 Ca+ per rhodopsin in

rods from in vivo dark-adapted retinas. However, rods ex-

tracted from retinas which had been removed prior to light

exposure and subsequently bleached contained only slightly

more than .25 Ca+ per rhodopsin. These results indicate

that upon light exposure in vivo, Ca++ is taken up by rods

probably from the pigment epithelium. Since she used bleach-

ing levels which were very high, it is not clear how (or if)

this relates to the Ca hyperpolarization theory.

It appears that the presence of trace elements and their

function in photoreceptor cells is unclear. This chapter

contained two different methods of determining trace element

content, radioactive tracers and atomic absorption spectro-

scopy. Most researchers using radioactive tracers have to

sonicate their samples and then claim that upon resealing the

vesicles are comparable to intact disc membranes. However,

no one has presented any evidence that this is the case. The

one study which did not sonicate their discs obtained vastly

different results from those authors who did.

The other technique, atomic absorption spectroscopy,

avoids the problem of sonication. However, the results ob-

tained with this method do not agree any better than those

obtained with radioactive materials. An explanation may be

indicated in two statements by Szuts and Cone: "Even though

our procedure was designed to minimize it, the unequal di-

vision of residual contamination by non-outer segment debris

between test and control aliquots may have been the cause of

the observed variability between experiments" and "in view

of the variabilities we observed, it seems likely that an en-

tirely different experimental approach will be needed to a-

chieve such a low detection limit" (p 206).25 However, the

problem with atomic absorption spectroscopy is not one of

detection limits, but is instead, one of not being able to

detect small changes in a relatively high content.



The technique of particle induced X-ray emission (PIXE)

depends on the high cross section for ionization and subse-

quent X-ray emission after excitation by a heavy, charged

particle incident upon the target element. The various theo-

ries which interpret this process will be discussed in this

chapter. (The next chapter describes the PIXE technique in


The simplest models assume the production of an inner

shell electron vacancy is the result of the Coulombic inter-

action between the incident particle and the bound electron.

Three theories have been formulated to explain this inter-

action: the plane wave Born approximation (PWBA), the impulse

approximation (also called the binary encounter approximation
or BEA), and the impact parameter method. The first two

are valid at high energies (incident particle energies much

higher than the electron's binding energy) and are more useful

for the understanding of PIXE. The impact parameter method

was developed for adiabatic scattering (incident energy close

to binding energy) and will not be considered in this work.

It can be shown that the validity criteria for both the

Born approximation and the impulse approximation are similar

for K shell ionization by protons and can be expressed as:

(V.1) Ep/Uk > 24

Where Uk = electron's K shell ionization energy
and Ep = proton energy.33

This will be considered the "high energy" region and will

always be valid for the experiments performed in this researCh.

In first order perturbation theory, the cross section for

a transition is proportional to the absolute square of the

matrix element of the perturbing potential:

(V.2) d n (Rr) V n (R,r) drdR
d(V.2) n n
where V = interaction potential

R = position vector of the incident particle

r = position vector of the target particle

Sn' = wave function of the final quantum mechanical state

n = wave function of the initial quantum mechanical


and d = differential cross section per solid angle.34

Using the PWBA with a Coulombic interaction between the inci-

dent particle and electron, the differential cross section can

be written:

(V.3) da Mv' in (r) ze exp [/h(p-p) R]
dP2 4-h v _v,

in (r) drdR

where LM = reduced mass of the atom particle system

e = electron charge

v = speed of relative motion before collision

v' = speed of relative motion after collision

p = momentum of relative motion before collision

p' = momentum of relative motion after collision

'n' = final wave function of the electron

n = initial wave function of the electron

z = projectile charge

h Planck's constant divided by 2n

i= ir

and the remaining symbols are as previously defined. (In

going from equation V.2 to equation V.3, a choice of "initial"

and "final" states has been made. This is equivalent to

making a choice of the Hamiltonion of the system."

Approximating the initial and final states by hydrogenic

wave functions and neglecting all screening effects, the authors

show that an approximate expression for K shell ionization is:

220, a2 T4
(V.4) = 4 a2

where k MZZ(Rydberg)

S= total ionization cross section

M = mass of incident particle

E = energy of incident particle

m = electron mass

a = first Bohr radius of hydrogen

Z = atomic number of target atom

and the other symbols were previously defined.34 Thus, nk is

a dimensionless number of depending upon the energy of the

incident ion and the energy of the electron.

In this theory, the cross section varies as the fourth

power of the incident ion energy. Thus, the loss of energy

by the beam as it passes through the target material will re-

sult in a decrease in the cross section for K shell ionization.

(As will be discussed in the next chapter, this should be

considered before PIXE targets are prepared.)

Another important prediction of equation V.4 is a smooth

variation of the cross section as the energy changes. It

also indicates the cross section is proportional to the in-

verse twelfth power of the target atom's atomic number. Thus,

the K shell ionization cross section by heavy, charged par-

ticles is comparable in magnitude for atoms of similar atomic

number and will decrease rapidly with increasing atomic number.

The binary encounter approximation considers the target

electron as a free particle. The role of the nucleus is

solely to establish the momentum distribution of the target

electron. The cross section for the interaction of the inci-

dent particle with a free electron is then summed over all

moment compatible with an energy exchange which is greater

than the ionization energy of the inner shell electron. Fi-

nally, it is weighted by the momentum distribution of the

bound electron states. The .derivation is somewhat lengthy

and complicated and the resulting equations are lengthy and

broken up into various integration regions dependent upon

theoretical parameters. Therefore, the derivation will only

be outlined and some comments made concerning the results.

The derivation is presented in the review article by Vriens35

and thoroughly discussed by Garcia et al.3

For a heavy, charged particle incident upon an electron,

the cross section per unit momentum transfer can be written:

TT z2 e4
(V.5) a dp = 8 2p dp (the familiar Rutherford scatter-
p p ing formula)

where p = momentum transferred from incident particle

to target electron

dp = incremental momentum transferred

and the other symbols are consistent with previous usage.

An electron will be ejected if the amount of energy trans-

ferred is greater than the ionization energy. Therefore:


(V.6) Qk = oE dE

where Qk = ionization cross section for the K shell

AEmax = maximum energy which can be transferred to

the target electron.

Equation V.6 is valid only for one electron speed. In

order to obtain the proper result Qk must be integrated over

all allowed momentum values for the electron.36 Thus:

(V.7) k (v) = Nk (vv ) F )dv
k 1 2 k 2 2

where ok( ) = total ionization cross section for the

K shell

Nk = number of equivalent electrons having

binding energy Uk

v = electron speed

fk(v) = speed distribution function of the electron.

And Qk(v ,v ) = ionization cross section as before except

the explicit dependence upon v and v is
1 2

The speed distribution function fk(v ) can be determined

"classically" from the micro-canonical ensemble using hydro-

gen energy values or quantum mechanically utilizing hydrogenic

wave functions. In either case the result is:

32 5 V2
(V.8) f (v ) -3 v5 V2

(v2 + v2)4
2 0
where v = ( -)
o0 m

and Et = total energy of the hydrogenic electron.36

Utilizing equations V.7 and V.8 Garcia et al. obtained

the total K shell ionization cross section.36 Even though

the result is somewhat more lengthy than the PWBA result, it

also is a smooth function of incident particle energy and de-

creases with increasing atomic number of the target atom.

Clearly the two approximations were found using different

approaches and would not be expected to agree very well. They

can be easily compared if the expressions are properly scaled

so that U2 ok/Z is plotted versus Em/MUk. This plot is shown

in Figure 2. The curves plotted for the PWBA contain an ef-

fective charge to account for electronic screening of the nu-

cleus and an effective potential to account for the reduction

in binding energy of the K shell electron due to the outer

electrons. These modifications are incorporated into a simple







0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2




Curve A P~ 7B (no screen-
Curve B IFJBA (screening)
Curve C BEA

I I f i

- J I r 1

parameter which is equal to one if all screening is ignored

and decreases as screening is added. Curve A represents no

screening and curve B a "reasonable" value of screening.

The agreement between curve A and curve C (BEA) is reasonably

good and if screening is added the agreement is improved sig-


Many experiments have been performed measuring K shell

ionization cross section and comparing the results with theo-

retical predictions. A summary of the experimental work done

in the MeV range up to January 1976 is found in Table I of

Johansson and Johansson.37 Table I of this dissertation

contains references to recent experiments which studied K

shell ionization due to protons in the MeV energy range. The

comparison between experiment and theory can be summarized

by commenting that the theories are within 5% to 30% of ex-

perimental results in the energy range used in this study.

This indicates that theoretical cross sections are probably

not valid for absolute determinations, if accurate elemental

analysis is desired.

Up to this point only the ionization cross section has

been considered. However, once an inner shell vacancy is

produced the ionized atom may rearrange itself by other

mechanisms, the emission of an Auger electron or a Coster-

Kronig transition. Coster-Kronig transitions do not occur
for K shell vacancies3 and will not be considered. Since

Auger electrons are not detected in PIXE it is important to

consider the relationship of the X-ray production cross section

to the ionization cross section, which for a particular X-ray



Author and Reference (a) Year
Liebert et al.39 1973

Lear and Gray4 1973

Bearse et al.4 1973
Gray et al.42 1973

Akselsson and Johansson 1974

Criswell and Gray44 1974

Ishii et al.45 1974

Tawara et al.46 1974

Khelit and Gray47 1975

Twara et al. 1976
Mlilazzo and Riccobono 1976

Wilson et al.50 1977

Kover51 1977
Tawara and Huchiya5 1977
Bissinger et al.5 1976

(a) All references are to reports of proton induced cross

line is as follows:

(V.9) a = .wb
xp 1

where a = X-ray production cross section
a. = ionization cross section

w = fluoresence yield (probability that the

ion emits an X-ray)

and b = relative intensity of possible transitions

to fill the inner shell vacancy.

The theory of fluoresence yields will not be discussed

except to point out that all theories have predicted that the

fluoresence yield is a constant for a particular K shell ioniza-

tion within the projectile energy range utilized by PIXE.5455

The importance of this will be clear after the next chapter.

Table II gives fluoresence yields for ions of interest in

PIXE spectra. It can be seen that they vary widely between

light and heavy elements. The difference between elements

studied in this work is approximately an order of magnitude.

It is also obvious that the higher production cross sections

for light elements aresomewhat negated by lower fluoresence


The other consideration in equation V.9 is the probability that

the transition in question will occur to fill the vacancy.

These probabilities have been extensively studied both theo-

retically and experimentally and are known to be constant for

a particular inner shell vacancy.55 (Two of the previous
references, Bambynek et al.54 and Salem e al.55 are excellent
references, Bambynek et al. and Salem et al. are excellent




P .0604
S .0761
C1 .0942
Ar .115
K .138
Ca .163
Sc .190
Ti .219
V .250
Cr .282
Mn .314
Fe .347
Co .381
Ni .414
Cu .445
Zn .479
Ga .510
Ge .540
As .567
Se .596
Br .622
Kr .646
Rb .669
Sr .691
Y .711
Zr .730
Nb .748
Mo .764

(a) All fluoresence yields are taken from reference 56.

review papers on fluoresence yields and transition probabili-

ties respectively.)

An area which could be extensively discussed in this

chapter is the energy dependence of the spectra obtained in

X-ray emission studies. It is obviously important that the

researcher be familiar with the energy levels and selection

rules for atomic transitions.

Inner shells of atoms tend to strongly reflect the inter-

action of the nucleus with the inner shell electron. (This

in part justifies the use of hydrogenic wave functions in

describing inner shell processes.) As the nuclear charge in-

creases, the nucleus binds the inner shell electron "tighter".

The electrons also repulse each other, the outer shell elec-

trons "pushing" the inner shell electrons "closer" to the

nucleus. Both of these effects spread out the allowed energy

levels as the atomic number increases. This results in higher

energy transitions for heavier atoms. Thus, the energy of a

particular transition varies systematically with the atomic

number, and can be used as a means of detecting and distin-

guishing among the elements.

The X-ray spectrum must follow the quantum mechanical

selection rules for electric dipole transitions. (Electric

quadrupole and magnetic dipole transition probabilities are

lower by a factor (Zeff/137)257 Thus, electric dipole tran-

sitions will contribute the most intense lines for the elements

in this study.) These selection rules eliminate most tran-

sitions and give rise to relatively simple X-ray spectra.

In spectroscopic notation, transitions to the lower (n=l)

shell are referred to as K transitions. A transition to the

next highest shell (n=2) is an L transition. (Transitions to

other shells are lettered, but are not of interest in this

paper.) This study predominantly involves K shell transitions.

(One L line will be considered.) The theories of L shell

vacancy production and fluoresence yields are much more com-

plicated and not as well studied as those of K shells.

The K shell transitions are divided into two groups, K

and K, transitions. KI transitions represent electrons

falling from the n=2 (L shell) to the n=l (K shell). K

transitions include all other X-rays emitted when a higher

shell electron fills a K shell vacancy. K and K energy

levels for transitions in the commonly used PIXE energy range

are shown in Table III. K and K transitions can be sub-

divided into transitions with different energies. This is

not normally done in PIXE work because of limitations in de-

tector resolution.

The X-ray lines which are commonly used in PIXE are shown

in Figure 3. Electric dipole transitions are discussed in

most graduate physics quantum mechanics texts and also the
classic work of Condon and Shortley. The only require-

ment of X-ray transition probabilities and energies is that

they are consistent for a particular inner shell vacancy.

This has been shown to be true for K shell vacancies of all

atoms studied.56




P 2.015 2.142
S 2.308 2.468
C1 2.622 2.817
Ar 2.957 3.191
K 3.313 3.589
Ca 3.691 4.012
Sc 4.090 4.459
Ti 4.510 4.931
V 4.952 5.427
Cr 5.414 5.947
Mn 5.898 6.492
Fe 6.403 7.059
Co 6.930 7.649
Ni 7.477 8.265
Cu 8.047 8.907
Zn 8.638 9.572
Ga 9.251 10.263
Ge 9.885 10.984
As 10.543 11.729
Se 11.221 12.501
Br 11.923 13.296
Kr 12.648 14.120
Rb 13.394 14.971
Sr 14.164 15.349
Y 14.957 16.754
Zr 15.774 17.687
Nb 16.614 18.647
Mo 17.478 19.633

(a) Energies are in keV


3 d 3/2

3s12 --

0 ^/

2P3/2 -



Is1/2 Ka

Ka2 Ka3 KP,

K/3 K?5






In 1970, a study was published by Johansson et al.59

demonstrating the feasibility of combining X-ray excitation

by heavy, charged particles with an energy dispersive X-ray

detector as a sensitive multi-elemental analysis technique.

Since that beginning, the field has grown such that during

a 1978 small accelerator conference in Denton, Texas, 19

papers were presented utilizing particle induced X-ray emis-

sion.60 This chapter describes the PIXE technique and attempts

to indicate why it is so useful. (The best review article

available on the PIXE technique is by Johansson and Johansson37

and is the source of most of the material in this chapter.)

PIXE requires a beam of charged particles, a target, and

an X-ray detector. Typically the ion beam is obtained from a

Van de Graaff accelerator; however, other types of accelerators

have been successfully used. The target which consists of

material to be analyzed is placed in the beam's path. One of

the effects of the beam upon the target is to remove inner

shell atomic electrons. The theories describing this process were

discussed in the previous chapter. The ionized atom can re-

arrange by emitting characteristic X-rays which are detected

in a suitable X-ray detector. As previously mentioned, the

energy of the X-ray allows the detection of the element re-

sponsible for its emission.

Target preparation is an important part in any PIXE experi-

ment. A researcher wants to alter the target as little as

possible; however, the constraints imposed by the PIXE tech-

nique usually require some preparation.

The most obvious requirement for most PIXE targets is

that they must be stable in a vacuum. (It is possible to

perform PIXE with the target in air utilizing an external

beam.) This means that biological samples must be subjected

to some method of water removal. Ideally a drying procedure

would be rapid, non-contaminating, simple, and remove only

water from the target. Air drying, freeze-drying, vacuum

substitution and ashing (both low and high temperature) are

most often used. Air drying is the simplest choice, but the

length of time required can be a disadvantage. However, it

probably remains the most common method of water removal from

biological samples. The other methods require appropriate

equipment, are more involved, and can remove volatile ele-

ments from the sample.

The next requirement for target preparation is that

both the absorption of emitted X-rays in the target and energy

losses of the incident beam due to the target be considered.

In the preceding chapter, it was shown that the cross section

for ionization depends upon the incident particle energy.

Since the beam particles lose energy during the interaction

with outer shell electrons,the cross section would depend

upon the target thickness. Fortunately, this can be ignored

if target thickness is no greater than 1 milligram per square

centimeter. This is often stated in papers reporting PIXE

experiments without reference to any calculations. It is

shown to be the case in Appendix A. The absorption of X-rays,

particularly softer X-rays emitted from elements which are of

biological interest, is a different problem. Studies on X-ray

absorption indicate that a target thickness of one milligram

per square centimeter would require a 20% correction for light

trace elements of biological interest (e.g., S, Ca and K).

Many researchers attempt to make very thin targets and sub-

sequently ignore X-ray absorption in the sample, while others

only look at elements with high energy X-rays for which the

absorption is not as great. However, for targets of typical

PIXE thickness (milligrams per square centimeter) and elements

in the range Z < 20 some type of correction must be made for

accurate results.

After the water has been removed from the target material

it must be formed into a thin target. For most materials, this

requires the use of a backing for strength. Commonly used

backings and the advantages and disadvantages of each are given
in Table 6 of Johansson and Johansson.37 If the target materi-

al is a serum or in solution, it may be placed directly upon

the backing; however, this often leads to targets which are

not uniform. If the data reduction depends upon target uni-

formity this can be a serious problem. Tissue samples prob-

ably present the most difficulty with target preparation and

numerous methods have been employed to overcome these.37 The

final type of target is one in which a strong backing is used

to filter the desired target materials from a fluid. This is

done in aerosol samples and in water contamination studies.

Since the fluid passes through the filter, the target is im-

mediately ready for analysis.

The above discussion should indicate some of the charac-

teristics of a good backing material. It should be thin but

strong, easy to handle, resistant to the target (some target

materials destroy the backing), stable in an ion beam, and

contain a minimum of contaminants. The diversity of target

preparation methods and backing materials used attest to the

fact that each problem must be considered individually.

Emitted X-rays can be detected using any type of X-ray

detector. However, the use of a wavelength dispersive detector

defeats the multi-elemental nature of PIXE or would require

very long runs to obtain a complete spectrum. Therefore, al-

most all PIXE work is done with energy dispersive detectors

and predominantly with lithium-drifted silicon Si(Li) detectors.

These detectors have an energy resolution of approximately

150-180 electron volts (when new) with a 30 square millimeter

detector area.

Since the ultimate sensitivity for PIXE will be strongly

influenced by signals which do not result from the desired

X-ray transitions, the sources of X-ray background in PIXE

experiments will be briefly considered.

The background spectra can be divided into three parts.

One part is due to bremsstrahlung from the incident particles.

Most of the high energy background is due to this process.

(For example, Figure 5 is the X-ray spectrum of a thin formvar

film which contains no elements with X-ray transitions in the

energy range shown. The X-rays above channel number 200 are

almost solely attributable to this source.) The largest por-

tion of the low energy background is due to bremsstrahlung

from secondary electrons. This results from the interaction

of electrons removed from the target atoms with the matrix.

Therefore, this portion of the background is due to the same

process which produces characteristic X-rays, the production
of electron vacancies.6

The third process contributing to the X-ray background

is due to nuclear excitation of the target atoms by the inci-

dent beam, emitting y-rays which result in Compton scattering

in the detector. This process is hard to predict theoretically

and is often ignored by researchers who use low energy beams.

However, it can be modeled and is almost constant with X-ray

energy in the energy range of Si(Li) detectors. It can also

be decreased by an appropriate choice of beam energy and par-

ticles (i.e., protons and low energies).62

It is clear from the previous discussion that the choice

of beam energy can effect PIXE background and hence sensitiv-

ity. Also, as Figure 2 indicates, the cross section for

inner shell ionization can be maximized by an appropriate

choice of particle energy. Thus, the choice of a beam energy

is a factor which should be given considerable thought in a

PIXE experiment.

In addition to beam energy, sensitivity also depends upon

the amount of charge incident upon the target. This can be

seen considering the relationship between peak area and back-

ground under the peak. The criterion normally used for sta-

tistical significance is:

N 3 / N

where Np = number of X-ray counts in the peak

and NB = number of X-ray counts in the background

under the same peak.37

Therefore, an increase in total charge incident upon the

target results in an increased ability to detect an X-ray


Using this criterion and conservatively estimating param-

eters for detector resolution (165 electron volts), solid

angle (.003 x 4r), target thickness (.1 milligram per square

centimeter), and total charge (10 microcoulombs), Johansson

and Johansson predict minimum detectability curves for an

"organic" target.37 These curves are shown for two energies

in Figure 4. They indicate that two MeV is a good choice of

energy for biological work. (This also is a low enough

energy to avoid y-rays from the prevalent biological matrix

elements, C and 0.)

Another factor which must be taken into account in PIXE

is the beam uniformity. If a nonuniform beam is incident

upon a nonuniform target nonuniformm in directions perpen-

dicular to the beam), erroneous data may be obtained. How-

ever, this can be avoided if either of the two is uniform.

Most PIXE researchers produce uniform beams through the use

of a beam sweep, diffuser foils, or beam defocusing.37

One of the most difficult problems in obtaining quan-

titative data is calibration. As previously mentioned, theo-

retical cross sections are not suitable if accurate results


- 2

^ ()

uo!ii ju@Duoo


are wanted. Instead, most researchers have used one of two

calibration methods. One is the incorporation of an "internal"

standard. In this procedure a known amount of an element

not present in the target material is added. The amount of

one element in the target is now known and can be used to

determine the amounts of other elements. The other method,

which is more common, is the use of "external" standards.

Standards for which the composition is known are subjected

to a known amount of charge. The number of counts under each

X-ray peak yields an absolute "system efficiency curve" which

contains the probability of an X-ray being detected per atom

per unit charge. The use of "internal" standards also incor-

porates a "system efficiency curve," but it is relative and

relates the system efficiency of each element to the others.

An interesting variation of a relative efficiency curve and

external standards is used by Lear et al.63 In this method

a relative efficiency curve is determined utilizing external

standards while the backscattered particles are monitored.

The number of backscattered particles is then used to relate

the target to standards.

The most difficult problem in PIXE experiments is spec-

trum analysis. Most samples of interest can contain many

peaks upon a background due to effects previously mentioned.

The most common procedure is to computer fit the X-ray peaks

to Gaussian curves. This can be accomplished relatively

easily unless interfering lines are encountered. The hand-

ling of overlapping peaks can present considerable difficulty

and is handled in various ways by PIXE researchers.

There are numerous computer programs which are used to

fit PIXE spectra. Most are modified nuclear physics programs

and are not able to properly model PIXE background. Absorption

by the target is considered by only one of the fitting pro-

grams. Absorption by the system components, windows to the

X-ray detector, air gaps and external absorbers, are already

included in the "system efficiency curves." Thus, if data are

wanted on low Z elements in biological targets, the limita-

tion will probably be in the fitting program.

There are many other aspects of PIXE which have not been

mentioned in this chapter. Some are not pertinent to this

study, others will be considered in subsequent chapters deal-

ing with experimental procedures, and the remainder were not

considered important enough for inclusion.



The previous chapters have dealt with a diversity of

subjects. Initially, the problem of transduction in photo-

receptors was outlined, including the relevant experiments,

and the dominant theory. Experimental attempts to verify

the Ca+ transport theory were mentioned. The subsequent

chapters tried to indicate that PIXE offered promise as a

technique which could be utilized in this research, and included

a brief theoretical foundation and description of the technique.

(There seems to be no doubt that PIXE can have a tremendous

impact upon clinical work; however, there have been few

attempts to apply it to "fundamental" biophysical problems.)

The remaining chapters will describe the experiment

which attempted to utilize PIXE in the determination of ionic

content to disc membranes and the detection of changes due to

light exposure, and the results of this experiment.

Bovine retinas were purchased from Hormel (Austin, Min-

nesota). The procedure used by Hormel to collect and freeze

the retinas is described in Appendix B. The retinas, contained

in vials of 50, were shipped in an insulated container and

placed between layers of dry ice. The shipments were made

by air freight. When each order was received, it contained

some dry ice. Therefore, it was felt that the retinas had

remained frozen during the shipment. They were then stored

in a freezer at -100C.

When disc membranes were needed for an experiment, they

were prepared using a separation procedure published by Smith

et al.64 This procedure is described in Appendix C. All

preparation was done in a darkroom using dim red light filtered

with a Kodak 1 filter. Prior to beginning the separation,

the retinas were allowed to thaw in the dark for approximately

two hours. Shorter thawing times reduced the yield signifi-


After separation, the intact disc membranes were suspended

in five milliliters (ml) of 57 sucrose solution. The solution

was divided into a number of equal parts (typically eight),

placed in small transparent vials, and stored on ice until


For light exposure, the vials containing discs were placed

on a counter at a predetermined location and subjected to a

predetermined amount of light. For each vial subjected to

light, a control was located at the same distance from the

light source, but was covered with an opaque object. For

the entire study there are equal numbers of control and ex-

posed tubes. The levels of exposure ranged from approximately

.1% bleaching to nearly 100% bleaching. For the calculation

of bleaching levels, see Appendix D.

Immediately after being exposed to light (or used as a

control), the vials were placed in a Beckman "Microfuge" and

spun at 11,900 revolutions per minute for one or five minutes.

(The force during this time was 8874 xg.) This sedimented

most of the disc membranes. After the supernatant was re-

moved, the sediment was pipetted onto previously prepared

thin formvar films. The top of the sediment was easily dis-

tinguishable from the sucrose medium and was completely re-

moved. Therefore, the amount of sucrose medium remaining in

the sediment should be proportional to the amount of sediment

in the target. (This has been assumed by previous researchers

attempting to measure ionic content. This was originally a

necessary assumption for this work, however, as will be seen

later, is not required.)

Before target preparation, but on the same day, 2% form-

var backings were made. These were discarded if not used

that day. The choice of backing has been considered previously.

Formvar was chosen for this work because it can be made with

a lower level of contaminants than other easily produced

backings. The fact that formvar can be made very thin was

also a factor in its being used. The films were floated on

double distilled water and picked up on aluminum frames.

After the targets were prepared, they were placed in a

closed container to air dry. This required two or three days

depending upon the amount of material in the target. Blank

formvar backings remained with the targets at all times after

they were prepared. This was to ensure there were no contami-

nants due to external contamination.

Targets were then taken to Florida State University (FSU),

Tallahassee, Florida,where they were subjected to PIXE analy-

sis. The proton beam of five MeV was obtained from the FSU

FN accelerator. (This is clearly too high an energy for

optimizing sensitivity for the elements in this study. How-

ever, the most important limitation upon beam energy is what

can be obtained and the FSU FN accelerator does not run well

at lower energies.) The beam was made uniform by means of

a beam sweep and subsequently collimated by the use of a car-

bon collimator of diameter 4.5 millimeters located one meter

in front of the target chamber. The accumulated charge and

beam current were monitored by a Faraday cup located past

the target. Beam currents were kept low, approximately five

nanoamperes, in order to prevent pulse pile-ups in the X-ray

detector. Each target was analyzed twice, once for an accu-

mulated charge of one microcoulomb and then for an accumulated

charge of 20 microcoulombs. (The reason for this will be ex-

plained later in this chapter.)

The resulting X-rays were detected in a Si(Li) Kevex

detector with a measured full width at half maximum of 200

eV at 6.4 keV. Count rates were between three and five thou-

sand counts per second. The exit window from the PIXE chamber

was .0064mm mylar. The total air gap between the exit window

and detector window was seven millimeters. And the detector

window was.013 mm Be. The vacuum in the PIXE chamber was be-

low 10- torr; therefore, no correction for X-ray absorption in

the chamber was made. The target was located in a vertical

plane at an angle of 300 from the beam. The X-ray detector

was located 900 from the plane of the target.

The PIXE spectra were then fit with the computer code

HEX which was developed especially for PIXE spectra at FSU

aufan. n62
by Henry Kaufmann. HEX is one of the computer codes which

is able to account for target absorption of X-rays. It is

described in Appendix E.

In studies of disc membranes, it is preferable to relate

ionic composition to the amount of rhodopsin. This enables

ionic changes to be easily compared to the amount of rhodopsin

bleached. Normally this is accomplished by taking equal ali-

quots from the same suspension and assuming they contain equal

amounts of both contaminating material and rhodopsin. When

this work was begun, this procedure was attempted; however, it

was not possible to equalize rhodopsin in this manner. This

could easily be seen by looking at the "equal" aliquots.

There are methods used which tend to better yield "equal" ali-

quots; however, it was felt that a better method could be de-

vised through the use of the PIXE technique.

A second problem relating to the amount of material present

is the requirement for an accurate determination of the amount

of sample. This is normally done by weighing the dried sample.

However, the reproducibility of air drying may not be good

enough to use this as a base for determining ionic changes.

With both of the preceding problems in mind, plus a de-

sire to lower the PIXE calibration error as much as possible,

an attempt was made to find an element which could be used as

a true internal standard. It would be preferable if this ele-

ment were contained in rhodopsin, were not found in the rest

of the disc membrane nor the cytoplasmic space of the rod

outer segment and were not present in any chemicals used.

Fortuitously there seems to be such an element, sulfur.

It is well known that rhodopsin contains sulfurs as both

disulfides and sulfhydryls. The reactions of sulfhydryl

groups in rhodopsin have been studied by various research

groups. 65,66 The number of sulfurs in rhodopsin is still

debated with ten and twelve being the most accepted numbers.

Chen and Hubbell's number of ten sulfurs per rhodopsin will

be used in order to calculate ionic composition per rhodopsin.65

However, the number of sulfurs does not matter if only changes

in ionic composition are wanted.

It cannot be unequivocally stated that sulfur is not

contained in disc membranes in other than rhodopsin; however,

the lipid components are known and do not contain sulfur.

The remaining protein component in disc membranes is not known.

Rhodopsin is 85% of the membrane protein and it could be as-

sumed that most of the membrane sulfur would be in rhodopsin.

If there were other sulfurs in the disc membrane, the amount

of sulfur would then be a measure of the disc membranes in the

target material. And as before, the changes in ionic compo-

sition would not be effected. Regardless, for the remainder

of this study, it is assumed that all disc membrane sulfur

is in rhodopsin and that each rhodopsin contains ten sulfurs.

Because the X-ray production cross section for light ele-

ments is so large and they are usually present in high concen-

trations, the X-rays from these elements dominate the X-ray de-

tector in the configuration described earlier. For better

detection of heavier elements, an external absorber is placed

into the X-ray path. This absorbs most of the low energy

X-rays, but transmits most of the higher energy X-rays. For

each PIXE target, one run was made without an absorber yielding

predominately X-rays from light elements, and then a run was

made with the external absorber giving a larger contribution

from heavier elements.

A source of possible contamination by other than membrane

sulfur is sulfur contained in free amino acids. Taurine is a

commonly occurring sulfur containing amino acid which is found

in retinas. However, it has been shown to be absent in rod

outer segments.67

Another type of contamination which could affect the re-

sults is due to other cell organelles. Since the outer seg-

ment does not contain organelles other than disc membranes,

if the preparation were pure, this would not be a problem.

However, mitochondria are present in the inner segment and

could be present in the suspension after the outer segment

is broken off from the retina. Mitochondria contain large

amounts of Ca+ and Cu+ which could lead to erroneous values

for these elements.

An assay was performed to determine the presence of

mitochondria. The crude rod outer segment preparation which

was obtained after the first flotation on 45% sucrose was

assayed for succinate dehydrogenase. Succinate dehydrogenase

is an enzyme which is present in mitochondria and is commonly

used as a mitochondrial assay. The amount of crude rod outer

segment used in the assay was adjusted until activity was

easily determined. Then a comparable amount of final disc

membrane preparation was assayed. In this preparation, activity


was lower by a factor of 20. Therefore, it was concluded that

mitochondrial contamination was not significant. The assay

procedure is described in Appendix F.



Figure 5 is the X-ray spectrum of one of the formvar

controls stored with the PIXE targets. The contribution due

to bremsstrahlung, both incident particle and scattered elec-

tron, is clearly evident. There are no X-ray peaks present

due to elements having energy transitions in this range.

This is a typical formvar spectrum and indicates there was

no contamination during air drying and storage. The amount

of contamination on any formvar blank did not exceed .1% of

the target amount.

Figures 6, 7, and 8 show the X-ray spectra of chemicals

used in the separation procedure. Ficoll is seen to contain

small amounts of Ca, approximately 100 parts per million.

Sucrose contains Cu at an approximate level of two parts per

million. The KPO4 buffer is apparently free from any signif-

icant contamination.

Figure 9 is the X-ray spectrum of a disc membrane prep-

aration. It was run without an external absorber and clearly

shows the light elements P and S. Figure 10 is the same tar-

get run for a longer time with an external absorber. A com-

parison of Figures 9 and 10 indicates how the use of an exter-

nal absorber "brings out" the heavier elements with respect

to the lighter elements.

0 0



-1NNVH3/SiN o o





0 D

< LL
- I.














0 0






S N n o o

o 0






I* 1

C w

7 0


S/.O O

/ 0


L !

^ ------------------o M

0 0 0




: ." C -

:.". -E

'. X

0 C> 0 0
.. NN H/Si
SI.^ g

N '.'. *d C a

0 0 0

-N: o sn

The results for each target are given in Table IV. The

value given for each element is with respect to rhodopsin

making the previously mentioned assumptions on the S content

of rhodopsin. Included is the tube from which each target

was extracted and the amount of bleaching. No values are

given for P or K since they were contained in large amounts

in the chemicals used in the separation procedure. Therefore,

these values would be unreliable as the composition of disc

membranes. The range given for each value represents the

error calculated from HEX combined with an error which rep-

resents the choice of matrix composition. Both of these are

discussed in Appendix E.

Table V contains the amounts of the most prominent trace

elements detected in disc membrane preparations. These values

are determined by averaging over all of the control values.

The range given is the standard deviation of the data.

Even though the major reason for performing these experiments

was to determine the trace element composition of disc membranes,

the presence of other elements can be detected as well as the

absence of numerous commonly occurring biological trace metals.

In addition to the trace elements listed in Tables IV and V, the

amount of chlorine in each sample was determined. The mean value

and standard deviation for all 22 samples are .903 .455.

Since HEX looks for the presence of any element requested

and is able to quantify the minimum concentration required for

detection, it is possible to place an upper limit on the amounts

of elements not detected. This was done for manganese, molybdenum

and nickel and resulted in upper limits per rhodopsin of .007,
.006, and .0033 respectively.


CD CA ,- CD C) D Lto UO CD U71 WD. ?v) 0t "A c, CD 0 CA VA CAJ CD 0-
r--W 0-~ 00 r- WDu)t ,-LoL)LnUoL)t i '0(Dt
CD C) C) C:)C) (- DCDC:)CD C:) (D (DC) rDCD (D C) C:)C:) (D
CDC C)))C) C D0C) CD)0cDc) 0CD C:)C:)C:)CC) )CDC:)C:)

+1 +1 +1 +1 +1 +1 +1 +1 +I +1 +1 +1 +f +I A-1 +1 +1 +1 +1 +1 +1 +!

CA tD CA -zt to I CA CN MCA on N-- to In Ma0 n L CD r-- 0 r- c
G7% C) r- ULr) r- to0 to to r- ',D '.1 If) to al ) LCr L ,-~ Z) C:)CD-) CDr40CD CD CDC) C) (C) C C:)CD() () () C) C

N Nd In N o -4 NIA

000000 00 -It0
-Aodd dC) DC) od d-d d-d4
CD C) (D0 0CD CDCC) (D(DC:) (DC) C)

+1 +1 +1 +1 +1 +1 +1 +! +1 +1 +1 +f +1 +! +1 +1 +1 +! +1 +1 +4 +1

-d CD C- oi I'D --:I- If) U-) CD --d- '0 '0 NA Cl) Ci 0-- NA r- in) CD 03- U-)

+1 +1 +1 +1 +1 +1 +1 4! +1 +1 +1 41 +1 +1 +1 +1 4! +1 +1 +1 +1 +1

I'D o- CO \'.0 CD OT -oC00 '0 r- CA N CD od CA CO C-1 on to C-
to0 r- "n C0'.0 -ci- CD r- If) r- c-

N' r- NA N No N ON d- ,-D r- 0 -,t N o N INN -,t r-i CDi a- l C-i to

uIn if) IC) '0 uin r- tci In)- -,3- t o r-- if) '.o '0 '.o c' oo '.o CD CD CD CDCDC)(C) CD 0CDCC) CDC)(DCD (D CD DCD C) (DCD

+1 +1 41 +1 +1 +1 +1 +1 +! +1 +i +1 +1 +1 +1 +1 +1 +1 +1 4! +1 +1

r-4 If) r--4 \.D 073 C-T' -n Co '0 \00 C) NI NA ()0 G) a-' CD l Ifo ani "I
r-- M- to0 d- .-4 NA U) CD ci -ci- '-4 '-4 o-Li I'D C'.0N CO C:) O'\ r- C--
IC) -,t -4 '. it) If) -ti tc oni --t CO I) -cit If) -,t [C) If-) -.I -c . . .

!A 1-




.Of VA 01
m-y > 0-3

'd4 C)


,y-o oN v- ini to) r.-- C-i C) y- NA y- on Co -t [ to0 C) -A on oni
,A my AN V4 CAi oni ononC-o o n a










Q-) (T





Ca (a) .492 .153

Fe .0243 .0075

Cu .132 .060

Zn .0577 .010

(a) Calculated by using control values from Table IV.



The primary goals of this study were to determine the

concentration of trace elements in dark-adapted frozen bovine

retinal disc membranes and attempt to measure any changes due

to light absorption. A secondary aim was to demonstrate the

usefulness of the PIXE technique in fundamental biophysical

research. This chapter discusses each of these, including

comments upon possible metallo-enzymes which may be present

in disc membranes. In addition, ideas for future research

which continue the visual transduction problem and enter

new areas in which PIXE offers unique capabilities are men-


Table V shows the elemental content of disc membrane

targets as determined by this study. All of the elements have

less than a one to one stoichiometry with rhodopsin. However,

this does not preclude higher levels in vivo with some subsequent

loss during preparation.

The amount of calcium present in disc membranes agrees

well with reported values by other researchers. Szuts and Cone

concluded that the calcium content is 0.2 calcium per rhodopsin
within a factor of two, Hendriks et al. reported 11 calcium

per rhodopsin (using a separation medium containing ATP),2
and Hess reported 0.25 calcium per rhodopsin.32 These values

all refer to dark-adapted disc membranes.


Even though none of the elements are stoichiometric with

rhodopsin, the concentrations detected are high enough to in-

dicate a biological significance. The amount of calcium is

large enough to indicate it could be the transmitter in Hagins

and Yoshikami's theory. Another role has been suggested

for calcium by Hess.68 Since she determined that cal-

cium content in rod outer segments was greater after light-

adaptation in vivo, she feels that calcium may have a role in

light-dark adaptation. (As background light intensity is

increased, the threshold for visual excitation also increases.)

However, she did not describe a mechanism for a calcium de-

pendent adaptation. Another possible role for calcium would

be to modulate enzyme activity. It has been well established

that calcium concentration is an important factor in several en-
zymatic rates. Therefore, the calcium level might be controlled

solely as a means of regulating enzymatic activity. It is

also possible that such an enzyme mediated by calcium could be

involved in visual transduction or light-dark adaptation.

The trace element with the next highest concentration in

discs is copper. Hess had previously detected copper in intact

rod outer segments and has suggested it may be due to the
presence of superoxide dismutase. Superoxide dismutase is a

cytoplasmic enzyme containing two zinc and two copper ions

per molecule and is responsible for destroying superoxide

radicals. It probably doesn't contribute significantly to the

values in Table V since the copper values vary widely. If this

enzyme were present in large amounts, the zinc values would

vary with the copper values. In addition, the enzyme is unlikely

63 4

to be present in the discs and any cytoplasmic enzyme should be

lost during the washes after the rod outer segments are

bursted. A role for copper in the functioning of disc mem-

branes has not be suggested.

It may be recalled that the sucrose used contained

copper as a contaminant. A solution of 5% sucrose containing

100 parts per million of copper would be .08 millimolar (mM)

copper. Therefore a small part of the copper found may be

due to contamination. However, the reason for the remaining

copper detected is not presently understood.

The case for a membrane associated zinc binding protein

appears to be much better. As Table IV shows, zinc values

are relatively consistent, except for sample 6, and would

indicate that zinc is closely associated with the disc membrane.

A possible choice for this enzyme is carbonic anhydrase which

has a molecular weight of about 30,000 and contains one zinc

per molecule.69 If all the zinc present were due to carbonic

anhydrase, it would be present at a level of .14 rM (using

a rhodopsin concentration of 2.5 mM) and constitute approxi-

mately 4% of the membrane protein by weight. Carbonic anhy-

drase seems to be a good possibility because it maintains

intracellular pH levels.69 In a semi-closed system such as

discs or rod outer segments, a mechanism would be required to

maintain the internal pH. High fluxes of ions occur and the

pH changes must be constantly compensated. If this were not

the case, the pH could vary detrimentally from physiological


As previously mentioned, Tam et al. 27concluded there

could be a one to one stoichiometry between zinc and rhodopsin

in vivo, but they were unable to detect zinc in preparations

from bovine retinas which had previously been frozen. Comparison

of their results with those of this study is difficult. However,

the separation procedure used here is "gentler" and might be

expected to result in less zinc removal.

The concentration of iron is the lowest of the trace

elements listed in Table V. Iron in biological systems is

always bound to high molecular weight substances. It may

be present in either a soluble protein or a membrane bound

protein. There are many known iron containing membrane bound

proteins.6 Many of these function as electron carriers and

some of these are present in mitochondria. Since the mitochondria

assay showed highly reduced activity in the final preparation,

it is felt that mitochondrial proteins are not the major con-

tributor to the iron content of disc membranes. Due to the

fact that iron containing proteins are widespread, its presence

is not surprising and agrees with the results of Hess.6 How-

ever, it is not yet appropriate to suggest a likely reason for

its presence in disc membranes.

The amount of chlorine (which would be present as the

chloride anion) is interesting. As the standard deviation of

the data indicates, the chlorine content varies widely between

samples. This is expected, as it probably is not closely

associated with disc membranes. Instead, it would be present

in the rod outer segment cytoplasmic space. Chloride could have

a role in visual transduction, but this has not been postulated.

Since the possibility of a role for barium in visual

transduction has been reported,31 an attempt was made to place

an upper limit on the amount of barium present in disc mem-

branes. The details of this estimation are contained in

Appendix G and yield an upper limit of .02 barium atoms per


The absence of manganese, molybdenum, and nickel above

minimum detectability levels is significant because it shows

there are no proteins which contain these trace metals at

high concentrations in disc membranes. The upper limits

determined are still high enough to allow the presence of

minute amounts of a protein containing one of these metals

as a part of the 15% non-rhodopsin disc membrane protein.

If the samples made from each tube are considered sep-

arately and the amounts of each element averaged, in almost

every case the values are correlated better than those for all

the controls. This is true even if the data for those samples

exposed to light are included in the average for each tube.

Therefore, there is a significant difference among the disc

membrane solution in each tube. Since this experiment used

tissues from animals whose diets are not controlled, this

result should probably be expected. (In retrospect, one of

the two major weaknesses of this study was the lack of any

control over the animals including their diet, light exposure,

slaughter conditions and retina removal.) Since 50, or 100,

retinas are contained in a tube, it would be expected that

the targets coming from the same preparation would be more

homogeneous. The data indicate this is true.

Since it was felt that the values should be reproducible,

at least in the control targets made from each tube, the con-

clusion was that the handling after the solution was divided

could be contributing to the difference. The most obvious

step to consider was the final sedimentation in the Microfuge

spin which could have caused unequal losses in trace elements.

Therefore, the spin time was lowered from five minutes for

tubes A and B to one minute for tubes C and D. This resulted

in zinc values which were relatively consistent in the samples

made from tubes C and D and some improvement in the calcium,

copper, and iron values.

The fact that the zinc values were made reproducible was

taken as an indication that the method of calibration with

respect to sulfur was valid and that only an improved proce-

dure of handling the samples between light exposure and target

preparation was required. Unfortunately, the scope of this

study precluded an analysis of effects of various centrifu-

gations and external medium upon the presence of trace elements.
Recently some research has been performed in this area and

indicates that a single low speed centrifugation can cause up

to a 60% loss of calcium. Any significant loss of trace elements

during centrifugation is probably not very reproducible.

Previous attempts at detecting a light effect on calcium

efflux have all relied on a similar type of centrifugation.

However, in none of the reports of these studies is there an

indication that unequal loss of trace elements during centri-

fugation was considered. In the experiment performed by

Szuts and Cone, the pellet was "formed" approximately five

minutes after having been subjected to light. The authors

do not indicate that they considered this step to be a poten-

tial problem, but as previously stated ascribe the disparity

to an unequal division of contaminating material. That pos-

siblility was considered in this study; however, the consistent

ratio of zinc to sulfur seems to indicate that this is not the

case. Since the final centrifugation is performed with

phosphate buffer as the medium, it is felt that the differences

in concentrations of calcium, copper, and iron represent

unequal losses and not contamination from the medium. (This

was the other major problem encountered during this study.)

The other major goal of this study was to detect changes

in the previously mentioned ionic concentrations of disc mem-

branes due to light absorption. The amounts of calcium,

copper, and iron were not considered reproducible enough to

detect changes due to low bleaching levels. However, this

experiment could have detected calcium changes as reported by

Smith et al.,21 Mason et al.,22 and Hendriks et al.26 and

zinc changes as reported by Tam's group. Each of these studies

utilized high bleaching levels and the first three indicated

at least a one to one release of calcium per rhodopsin bleached.

These changes could have easily been seen in the targets which

were totally bleached; 11, 13, and 31. A possible reason for

not observing any changes in zinc has been mentioned. However,

it must be concluded that changes in calcium due to high bleach-

ing levels were not present under the conditions of this experi-

ment. This agrees well with later more controlled attempts to

measure calcium changes in disc membranes due to light absorp-

23 25
tion. This does not conclusively prove the calcium trans-

duction theory wrong. It has been previously mentioned that

the bleaching levels of bovine retinas obtained from Hormel

are not controlled and could be high enough so that rods have

been desensitized. That this is probably the case is indicated

by Smith et al. who state that approximately 25% of the rhodop-

sin is bleached.64 If true, it is possible that the calcium

transduction theory of Hagins and Yoshikami can not be tested

utilizing these slaughterhouse retinas.

One very significant result of this study has been that

it has demonstrated the advantage of using PIXE as a tool

for studying biological problems. This has also resulted in

numerous ideas for future research, some of which will be

briefly discussed.

A very important problem which should be studied is the

determination of the ionic composition of dark-adapted disc

membranes. For this purpose, animals (probably frogs or rats)

should be dark-adapted in vivo, the retinas removed in the

dark, and quick frozen. Then thin slices can be prepared

with an ultramicrotome, placed on formvar backings, and examined

with PIXE. In order to determine whether the slice has been

made through rod outer segments, each slice can be examined

under a light microscope. These targets can be made so thin

that they can be subjected to Rutherford backscattering analysis

to determine the low Z composition not obtainable with PIXE.

This can be done at the same time characteristic X-rays are

being detected for PIXE analysis. Since the matrix responsible

for X-ray self absorption is made up of the low Z elements,

the matrix composition can be included in the computer code

yielding much more accurate values for self absorption. In

addition, the combined use of PIXE and Rutherford backscattering

analysis allows the entire elemental composition, except for

hydrogen, to be found.

As a modification of the previously mentioned study, the

animals should be light-adapted in vivo and disc membrane targets

prepared in the same manner. A comparison of the two experiments

would not test Hagins and Yoshikami's theory, but it would

clearly show if there are significant effects (which originate

outside the retinal rod) upon the ionic content of disc mem-

branes. This is of interest since the data of Hess32 indicate

that calcium may be released into rod outer segments from the

pigment epithelium during light-adaption. Each of these can

be done by comparing the sulfur content to that of the element

of interest as was done in this study.

Another study which could be easily done is to utilize

PIXE as a verification of the number of sulfurs per rhodopsin.

Disc membranes could be prepared by a suitable preparation

technique and analyzed with PIXE. Using the percentage of

rhodopsin in disc membranes and assuming that all the sulfurs

in disc membranes are contained in rhodopsin will yield the

number of sulfurs per rhodopsin.

As a result of this research, it is obvious to this author

that any reasonable attempt to measure calcium changes in disc

membranes due to light absorption must begin with animals which

are known to be dark-adapted, must use a procedure for ob-

taining disc membranes which does not cause the loss of ionic

components, and should use a method of internal calibration

similar to the one developed in this study utilizing sulfur.

Since the theory of Hagins and Yoshikami predicts at least ten

calcium ions released per rhodopsin bleached, it is felt that

this experiment could definitively test the theory at low

bleaching levels (1%) if losses during preparation can be pre-


One promising clinically oriented study is to examine

the effect of zinc deficient diets upon both retinal degenera-

tion and retinal zinc content to include disc membrane and

pigment epithelium content. This is needed research as there

are indications that zinc deficient diets are related to the

most common retinal disorder, retinitis pigmentosa. (William

Dawson, Department of Ophtalmology, University of Florida,

Gainesville, private communication). Rats, whose diets are

easily controlled, could be used in this study with existing

physiological techniques used to monitor the extent of retinal

disorder and PIXE used to determine the corresponding trace

element content.

The previously suggested research is all related to the

work done in this study. However, it should be obvious that

the PIXE technique can be applied throughout many biological

research areas and seems to be uniquely suited for research in

some of these. One broad area suitable for PIXE is that of

protein-lipid-ion association. One of the most important

problems in membrane biophysics is how the membrane components

associate to form a stable membrane. In order to study this

association, vesicles could be formed from phospholipid, a

membrane protein which contains sulfur, and various ions. If

the components are properly mixed, they will form membranes

which can be shown to have certain physiological functions and

can be separated from the medium by centrifugation.71 Targets

could be formed from these vesicles and analyzed by PIXE.

The amount of each component can be determined, since the phos-

pholipid content is given by the amount of phosphorus, the

protein component is given by the amount of sulfur, and the

ionic content is determined directly. Available components

could be varied among the phospholipids, numerous membrane

proteins which have been extracted, and ions which are of

interest in membrane formation and stability. This type of

study could yield valuable information concerning the types

of association required for membrane formation and stability.



The stopping cross section can be written as:

dE -Nc

where dE
dwhe = energy loss per unit distance

N = number of stopping atoms per unit volume
and c = stopping cross section per atom.7

For oxygen an empirical curve for e is given by

S= (1.92/E ) [in (-) + 5.12]10-15 ev-cm2/atom

where E = proton energy in MeV.

For 5 iHeV protons c = 1.79 x 10-15 ev-cm2/atom.

If a density of one gram per cubic centimeter is used
for biological samples then for oxygen, N 1 3.75 x 1022 atoms
per cubic centimeter. Then

= -67 MeV/cm
A biological target .1 milligrams per square centimeter would
be 10-4 centimeters thick and the energy loss in traversing
this target would be .0067 MeV.

If a more realistic composition were assumed, including
carbon and hydrogen, e would be slightly lower. Thus, it
is concluded that the change in production cross section can
be ignored in this study.



Within 20 minutes after the animal is slaughtered the
whole eyes are removed and placed in a container of wet
ice. After 50 are collected they are taken to a darkroom
where the retinas are removed under red light. Each retina
is then placed in a vial which is on dry ice, rapidly freezing
the retina. The vials are filled with 50 retinas, wrapped
in aluminum foil and transferred to freezer storage. The
collection of 50 eyes takes approximately 15 minutes as does
the removal of 50 retinas (John Schmitz, Hormal and Co.,
Austin, Minnesota, private communication).



Fifty bovine retinas were defrosted and shaken for
three-five minutes in 45% sucrose, .1 Molar (N) phosphate
buffer (pH 7). The suspension was equally divided and
spun at 25,000 revolutions per minute (rpm) in an SW-27
rotor. The float material from each tube was pipetted
into a Sorvall centrifuge tube. The tubes were filled
with 1 M-phosphate buffer and spun in a Sorvall SS-34
rotor for 20 minutes at 15,000 rpm. The supernatant was
decanted from each tube and the pellet suspended in 40%
sucrose, 1 i-phosphate buffer. This was layered with .1
M-phosphate buffer and spun at 25,000 rpm for one hour
in an SW-27 rotor. The isolated rod outer segments were
pipetted from the interface. They were then placed in
double distilled water in a Sorvall tube and spun at 17,500
rpm for 30 minutes. The resulting pellet was suspended in
5% Ficoll and left for 10-14 hours in order to burst the rod
plasma membrane. They were then spun in an SW-27 rotor for
two hours at 25,000 rpm. The intact discs were pipetted from
the surface and spun twice in 5% sucrose at 17,500 rpm for
30 minutes in a Sorvall tube.64 All centrifugations in Sor-
vall tubes were done in an SS-34 rotor.



Bleaching was done with a standard 40 watt bulb rated at
455 lumens. At 555 nanometers one lumen equals 5.72 x 10-4
watts73 which equals 1.47 x 1015 quanta per second. There-
fore, a 40 watt bulb yields 6.69 x 1017 quanta per second.
At one meter distance the quantal flux is 5.32 x 1012 quanta
per square centimeter per second.

Rhodopsin has a molecular extinction coefficient of
1.56x10-20 square meters.74Ignoring screening effects the ratio
of rhodopsin bleached per second at one meter will be .00083.
A one meter distance was used for low bleaching levels. When
higher levels were desired (above 1%) the samples were placed
closer to the light. Samples which were 100% bleached were
exposed to room lights for a minimum of ten minutes and no
coloration was observable. In addition to screening, the
absorption of the plastic vials was ignored.



HEX is a Fortran program specifically designed for anal-
ysis of PIXE spectra. It is a non-linear least squares fit
to the data utilizing numerous parameters. Some are deter-
mined previously and held fixed during the fit, and others
are allowed to vary. However, most of the free parameters
represent physical parameters obtained from models of the data
acquisition process.

The background model combines both Compton scattering
and bremsstrahlung. The parameters dealing with Compton
scattering are empirical and determined during the fit. The
fitting function for Compton scattering is the sum of a con-
stant and a slowly decaying exponential. The bremsstrahlung
model was determined from the physical processes and these
parameters have physical significance. It is modeled by a
decreasing exponential.

The peaks are fitted with modified Gaussians which reflect
the loss of detector resolution as X-ray energy increases.
Each element in the library is given a fixed K. to Kg intensity
ratio which has been experimentally determined. Thus one
linear parameter is used to represent each element detected
in the sample.

The remaining parameters treat the absorption problem. The
absorption external to the target is calculated from X-ray
absorption coefficients of the material through which the
X-rays must pass after leaving the target. For each system
configuration a transmission curve is calcualted. This curve
is then held fixed during each PIXE fit. As previously men-
tioned, a unique feature of HEX is its ability to calculate
and correct for self-absorption. This is done with an expo-
nential containing two parameters which are determined in the
least squares fit and can be related to the composition and
thickness of the target.

All parameters are determined simultaneously during the
Marquardt least squares fit and the linear parameters modi-
fying the Gaussians' heights are then used to determine the
amount of the appropriate element. HEX incorporates system
efficiency values determined by bombarding commercial stan-
dards. These are guaranteed to 5% accuracy.

Error analysis is done by HEX allowing 5% error for
standards and 2% for electronics. In addition, the fitting

error is included. It is also determined by HEX from the
error matrix.

Part of the input data to HEX is the percentages of
light elements in the matrix. This is held constant and used
to calculate self absorption, consistent with the parameter
for thickness, in the fitting procedure. Since the composition
of the matrix was not known it was estimated from known pro-
tein and lipid compositions. These values were checked by
increasing the oxygen content (oxygen was the heaviest element
in the matrix) until a total mass equal to the measured mass
of the target was obtained. The other extreme is known be-
cause it would be a matrix with no absorption. The differences
between absorptions calculated at these two extremes was
typically 20%. Assuming a 95% certainty leads to a 5% error
in choice of matrix composition. This was combined with the
error determined by HEX for sulfur. The ratio of each detected
element to sulfur was computed and the errors combined as
independent errors.

It is obvious that some of the errors associated with
fitting, standardization, electronics and absorption must
cancel if the ratio of two elements is used. This is because
the errors are not independent and should have been combined
differently. However, a realistic calculation of the inter-
dependence was not possible and errors are as indicated. Thus,
it is felt that the errors reported are probably too large.

It should be noted that HEX does not consider silicon
escape peaks nor X-ray absorption edges. The effect of these
omissions is thought to be very small; however, it is not
possible to determine if this is true without extensive experi-
mentation. Regardless of this, it is felt that HEX is the best
data reducing program in use in PIXE work.



The assay used measures the change in absorbance at 455
nanometers due to the reduction of K3Fe(CN)6 by succinate
dehydrogenase. The change in absorbance is then related to
the succinate dehydrogen se activity through the known molar
extinction coefficient.75

The method of Bradford was used to determine the protein
concentration./ This enabled the succinate dehydrogenase
activity to be determined per milligram of protein. The result
for crude rod outer segment preparation was 5.6 micromoles/
minute/milligram of protein.



The L lines of Ba which have energies between 4.451
and 5.995 KeV lie between the K lines of Ca and Fe. This region
of the spectra appeared void of peaks indicating no Ba was
detected. However, if Ba has a role in the visual process it
should be present in detectable amounts. Therefore, it is of
interest to determine an upper limit on the amount of Ba present.

Using Figure 4, it can be seen that the detection concen-
trations for Z=56 (using the L line curves) are approximately
five times higher than Z=24 (chromium, using K line curves).
Since the computer code HEX yields a minimum detection limit
for each element contained in the library, these two can be
combined to yield the desired Ba estimate. The minimum detection
limit for chromium was typically two nanograms per square
centimeter implying a limit for Ba of ten nanograms per square
centimeter. Using sulfur values as 3 x 103 nanograms per
square centimeter yields an upper limit of .02 Ba per rhodopsin.


1) Hecht, S., Physiology Review 17, 239 (1937).

2) Rushton, W.A.H., Journal of Physiology (London) 134, 11

3) Wald, G., in Enzymes: Units of Biological Structure and
Function, edited by T.P. Singer (Academic Press, New York,
1956), p.355.

4) Wald, G., in Light and Life, edited by W.D. McElroy and
B. Glass (Johns Hopkins Press, Baltimore, 1961), p. 724.

5) Encyclopedia Britannica 23, 60 (1971).

6) Wald, G., Science 119, 887 (1954).

7) Wald, G., Science 162, 230 (1968).

8) Mason, W.T., in Biochemistry and Physiology of Visual
Pigments, edited by H. Langer (Springer-Verlag, New York,
1973), p. 295.

9) Nir, I. and D.C. Pease, Experimental Eye Research 16,
173 (1973).

10) Blaise, J.K. and C.R. Worthington, Journal of Molecular
Biology 39, 417 (1969).

11) Young, R., in The Retina, edited by B.R. Straatsma
(University of California Press, Los Angeles, 1969), p. 177.

12) Hall, I.O., S.F. Basinger, and D. Bok, in Biochemistry and
Physiology of Visual Pigments, edited by H. Langer
(Springer-Verlag, New York, 1973), p. 319.

13) Stone, W.L. and E.A. Datz, Photochemistry and Photobiology
26, 79 (1977).

14) Rothschild, K.J., J.R. Andrew, W.J. De Grip, and H.E.
Sanley, Science 19, 1176 (1976).

15) Stubbs, G.W., B.J. Litman, and Y. Barenholtz, Biochemistry
15, 2766 (1976).

16) De Grip, W.J., S.L. Bonting, and F.J.M. Daemen, in
Biochemistry and Physiology of Visual Pigments, edited
by H. Longer (Springer-Verlag, New York, 1973), p. 29.

17) Wald, G., in Biochemistry and Physiology of Visual
Pigments edited by H. Langer (Springer-Verlag, New
York, 1973), p.l.

18) Yoshikami, S. and W.A. Ragins, in Biochemistry and
Physiology of Visual Pigments, edited by H. Langer
(Springer-Verlag, New York, 1973), p. 245.

19) Hagins, W.A. and S. Yoshikami, Experimental Eye Research
18, 299 (1974).

20) Hagins, W.A., in Annual Reviews of Biophysics and
Bioengineering, edited by L.J. Mullins (Annual Reviews,
Inc., Palo Alto, 1971), 1, p. 131.

21) Smith, H.G., Jr., R.S. Fager and B.J. Litman, Biochemistry
16, 1399 (1977).

22) Mason, W.T., R.S. Fager and E.W. Abrahamson, Nature 247,
562 (1974).

23) Sorbi, R.T. and A. Cavaggioni, Biochimica et Biophysica
Acta 394, 577 (1975).

24) Liebman, P.A., Investigative Opthalmology 13, 9 (1974).

25) Szuts, E.Z. and R.A. Cone, Biochimica et Biophysica
Acta 468, 194 (1977).

26) Hendriks, T., F.J.M. Daemen, and S.L. Bonting, Biochimica
et Biophysica Acta 345, 468 (1974).

27) Tam, S.W., K.E. Wilbur, and F.W. Wagner, Biochemical
and Biophysical Research Communications 72, 302 (1976).

28) Fukami, I., B.L. Vallee, and G. Wald, Nature 183, 28 (1959).

29) Neufeld, A.H., W.H. Miller, and M.W. Bitensky, Biochimica
et Biophysica Acta 266, 67 (1972).

30) Hemminki, K., Vision Research 15, 69 (1975).

31) Brown, K.T. and D.G. Flaming, Proceedings of the National
Academy of Sciences, USA 75, 1537 (1978).

32) Hess, H.H., Experimental Eye Research 21, 471 (1975).

33) Garcia, J.D., R.J. Fortner, and T.H. Kavanagh, Reviews
of Modern Physics 45, 111 (1973).

34) iIerzbacher, E. and H.W. Lewis, in IIandbuch der Physik,
edited by S. Flugge (Springer-Verlag, Berlin, 1958), 34,
p. 166.

35) Vriens, L., in Case Studies in Atomic Collision Physics I,
edited by E.W. McDaniel and M.R.C. McDowell (North-Holland,
Amsterdam, 1969), p.337.

36) Garcia, J.D., Physical Review A 1, 280 (1970).

37) Johansson, S.A.E. and T.B. Johansson, Nuclear Instruments
and Methods 137, 473 (1976).

38) Coster, D. and L. Kronig, Physica 2, 13 (1935).

39) Liebert, R.B., T. Zabel, D. Miljanic, H. Larson, V.
Valkovic, and G.C. Phillips, Physical Review AS, 2336 (1973).

40) Lear, R. and T.J. Gray, Physical Review A3, 2469 (1973).

41) Bearse, R.C., D.A. Close, J.J. Halanify, and C.J. Umbarger,
Physical Review A7, 1269 (1973).

42) Gray, T.J., R. Lear, R.J. Dexter, F.N. Schwettmann, and
K.C. Weimer, Thin Solid Films 19, 103 (1973).

43) Akselsson, R. and T.B. Johansson, Zeitschrift fur Physik
266, 245 (1974).

44) Criswell, T.L. and T.J. Gray, Physical Review A10, 1145

45) Ishii, K., S. Morita, H. Kaji, C.N. Hsu, and T. Shiokawa,
Physical Review A10, 374 (1974).

46) Tawara, H., K. Ishii, S. Morita, H. Kaji, C.N. Hsu, and
T. Shiokawa, Physical Review A9, 1617 (1974).

47) Khelil, N.A. and T.J. Gray, Physical Review All, 893, (1975).

48) Tawara, H., Y. Hachiga, K. Ishii, and S. Morita, Physical
Review A13, 572 (1976).

49) Hilazzo, M. and G. Riccobono, Physical Review A13, 578 (1976).

50) Wilson, S.R., J.R. Rowe, J.L. Duggan, and F.D. McDaniel,
Proceedings of the 4th Conference on Applications of Small
Accelerators, p. 186 (1976).

51) Kover, A., Nuclear Instruments and Methods 142, 5 (1977).

52) Tawara, H. and Y. Hachiya, 2nd International Conference
on Inner Shell Ionization Phenomena, Frieburg, Germany,
29 Mar-2 Ap 76.

53) Bissinger, G., J.M. Joyce, H.W. Kugel, Physical Review
A14, 1375 (1976).

54) Bambynek, W., B. Craseman, R.W. Fink, H.U. Freund, H.
Mark, C.D. Swift, R.E. Price, and P.V. Rao, Reviews of
Modern Physics 44, 716 (1972).

55) Salem, S.I., L. Panossian, and P.A. Krausse, Atomic Data
and Nuclear Data Tables 14, 91 (1974).

56) Mitchell, I.V. and J.F. Ziegler, in Ion Beam Handbook for
Material Analysis, edited by J.W. Mayer and E. Rimini
(Academic Press, New York, 1977) p. 311.

57) Merzbacher, E., Quantum Mechanics (John Wiley and Sons,
Inc., New York, 1961) p. 452.

58) Condon, E.U. and G.H. Shortley (Cambridge University
Press, Cambridge, 1964).

59) Johansson, T.B., R. Akselsson, and S.A.E. Johansson,
Nuclear Instruments and Methods 84, 141 (1970).

60) IEEE Transactions on Nuclear Science NS-26 (1979)

61) Folkmann, F., C. Gaarde, T. Huus, and K. Kemp, Nuclear
Instruments and Methods 116, 487 (1974).

62) Kaufmann, H.C. and R. Akselsson, Advances in X-ray
Analysis 18, 353 (1975).

63) Lear, R.D., H.A. Van Rinsvelt, and W.R. Adams, Advances
in X-Ray Analysis 19, 521 (1976).

64) Smith, H.G., G.W. Stubbs, and B.J. Litman, Experimental
Eye Research 20, 211 (1975).

65) Chen, Y.S. and W.L. Hubbell, Membrane Biochemistry 1, 107

66) De Grip, W.J., G.L.M. Van De Laar, F.J.M. Daemer, and S.L.
Bonting, Biochimica et Biophysica Acta 325, 315 (1973).

67) Pourcha, R.G., Experimental Eye Research 25, 119 (1977).

68) Hess, H. H., Federation Proceedings 35, 1949 (1976).

69) White, A., P. Handler, and E. L. Smith, Principles of
Biochemistry (McGraw-Hill, Inc., New York, 1973).

70) Daemen, F.J.M., P.P.M. Schnetkamp, T. Hendriks, and S.L.
Bonting in Vertebrate Photoreception, edited by H.B. Barlow
and P. Fatt (Academic Press, London, 1977), p. 29.

71) Jain, M.K., The Bimolecular Lipid Membrane (Van Nostrand
Reinhold Company, New York, 1972).

72) Whaling, W., in Handbuch der Physik, edited by S. Flugge
(Springer-Verlag, Berlin, 1958),34, p. 193.


73) Allen, C.J. and G.D. Rowe, in Handbook of Chemistry and
Physics, edited by R.C. Weast (CRC Press, Cleveland, 1976),
p. E-204.

74) Dartnall, II.J.A., in Handbook of Sensory Physiology,
edited by H.J.A. Dartnall (Springer-Verlag, Berlin,
1972), VII/2, p.122.

75) Veeger C., D.V. DerVartanian, and W.P. Zeylemaker, in
Methods in Enzymology, edited by J.M. Lowenstein (Academic
Press, New York, 1969), 13, p. 83.

76) Bradford, M.M., Analytical Biochemistry 72, 248 (1976).


Larry Don McCormick was born in Pocatello, Idaho,in

October, 1944. He attended high school in Callahan, Florida,

and was graduated in 1962. Subsequently, he completed an

Engineering Science degree at the United States Air Force

Academy and was commissioned in the United States Air Force

in 1966. In 1967, he married his remarkable wife, Cathy,

and began a graduate program in physics at the University of

Denver, Denver, Colorado. He received a Master of Science

degree from the University of Denver in 1968. In 1973, he

resigned his commission and entered graduate work in physics

at the University of Florida.

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of DoctoS of Phil sophy.

Henri A. Van Rinsvelt, Chairman
Professor of Physics

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

'\ .j '- -<_ ~-
F. Eugene Dunnam, Cochairman
Professor of Physics

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

For Arnold H. Nevis C
Professor of Physics and
Electrical Engineering

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

James W. Dufty "
professor of Pfysics

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

Robert J. Cohen
Associate Professor of

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor f Philos hy.

Douglas L. Smith
Associate Professor of Geology

This dissertation was submitted to the Graduate Faculty of
the Department of Physics and Astronomy in the College of
Liberal Arts and Sciences and to the Graduate Council, and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.

June, 1979

Dean, Graduate School

3 1262 08666 920 6
3 1262 08666 920 6

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E8SSI9OTG_2NHL46 INGEST_TIME 2012-03-09T01:17:02Z PACKAGE AA00003895_00001