Enhancing and inhibiting diabetic amyloid misfolding

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content 1

Enhancing and Inhibiting Diabetic Amyloid Misfolding

Thesis by
Alan Kiyoshi Okada

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
AND KECK SCHOOL OF MEDICINE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Degree of
DOCTOR OF PHILOSOPHY
(Integrative Biology and Disease)

Defended June 15
th
, 2016
Degree Conferral, August 2016

2

Dedication

Thank you to my parents, George and Sharon, for giving me the freedom to choose my own path
and the tools to succeed.
Thank you to my Sensei, Martin Katz, for showing me the way.
Thank you to my wife, Leah, for walking it with me.
Thank you to my dog, Odin. You’re the best, bud.

3

Acknowledgements

For your mentorship, and guidance, thank you Dr. Ralf Langen. I’ve learned a lot. After this
thesis, I promise, no more throwing in the kitchen sink when I write.
For teaching me science and how to lead, thank you Dr. Robert Chow.
To Mr. Kazuki Teranishi, this is half yours. Thank you for everything.
Dr. Jose Mario Isas, you’ve been a teacher and friend. Couldn’t have done it without you.
To everyone who worked with me on these and all the other projects that didn’t make it into the
thesis, especially Chenura Jayawickreme, Fleur Lobo, and Anise Applebaum, thank you.
To all the membranes of the Langen Lab, Drs. Mark Ambroso, Jobin Varkey, Balachandra
Hegde, Sahar Bedrood, Nitin Pandey, José Manuel Bravo Arredondo, Natalie Kegulian, as well
as Prabhavati B. Hegde, Sean Chung, Meixin Tao, and Rachel Lee, thanks for everything. It’s all
of you who made the last five years so memorable.
I would like to express my gratitude to my collaborators for their kindness and dedication,
including Drs. Cohen and Camarero.
To my committee, Drs. Siemer, Chen, Chow and Langen, thank you for your time and energy
spent helping me through from day one.
I also thank USC for nurturing a positive and collaborative scientific environment.

4

Contents

Dedication ....................................................................................................................................... 2
Acknowledgements ......................................................................................................................... 3
ABSTRACT .................................................................................................................................... 8
Chapter 1 Introduction .................................................................................................................. 10
1.1 Protein Misfolding in Human Disease ........................................................................... 10
1.2 Amyloid Structure .......................................................................................................... 13
1.3 Protein-Misfolding Kinetics ........................................................................................... 16
1.4 Islet Amyloid Misfolding as a Pathogenic Component in Type 2 Diabetes Mellitus .... 19
1.5 Protein Misfolding Pathways of Islet Amyloid Polypeptide: Membrane-Mediated
Misfolding ................................................................................................................................. 24
1.6 Approaches to Modifying Protein Misfolding Disease Progression .............................. 28
1.6.1 Targeting Protein Misfolding .................................................................................. 28
1.6.2 Small Molecule Inhibitors of IAPP......................................................................... 30
1.6.3 Proteins and Peptides as inhibitors of IAPP misfolding ......................................... 31
1.6.4 Chaperones as Inhibitors of IAPP Misfolding ........................................................ 32
1.7 REFERENCES ............................................................................................................... 33
Chapter 2 Diabetic Risk Factors Promote Islet Amyloid Polypeptide Misfolding by a Common,
Membrane-mediated Mechanism .................................................................................................. 47
ABSTRACT .............................................................................................................................. 48
2.1 INTRODUCTION .......................................................................................................... 49
2.2 RESULTS....................................................................................................................... 52
2.2.1 Phosphatidic Acid-LUVs accelerate IAPP misfolding ........................................... 52
2.2.2 IAPP transiently adopts an α-helical conformation with PA-LUVs before
transitioning to a β-sheet rich structure ................................................................................. 55
2.2.3 Electron microscopy of IAPP incubated with PA-LUVs. ...................................... 58
2.2.4 Oleic Acid-LUVs accelerate IAPP misfolding kinetics.......................................... 59
5

2.2.5 IAPP transiently adopts α-helical conformation with OA-LUVs before
transitioning into β-sheet structure ........................................................................................ 62
2.2.6 Electron microscopy of IAPP incubated with OA-LUVs ....................................... 63
2.2.7 Monobenzyl ester phthalate-LUVs accelerate IAPP misfolding kinetics ............... 64
2.2.8 Electron microscopy of IAPP incubated with MBzP-LUVs. ................................. 66
2.3 DISCUSSION ................................................................................................................ 68
2.4 REFERENCES ............................................................................................................... 70
2.5 ACKNOWLEDGEMENTS ........................................................................................... 77
2.6 AUTHOR CONTRIBUTION STATEMENT ............................................................... 77
2.7 COMPETING FINANCIAL INTERESTS STATEMENT ........................................... 78
2.8 ABBREVIATIONS ........................................................................................................ 78
2.9 MATERIALS AND METHODS ................................................................................... 79
Chapter 3 Cardiolipin Accelerates the Misfolding of Islet Amyloid Polypeptide ........................ 84
ABSTRACT .............................................................................................................................. 85
3.1 INTRODUCTION .......................................................................................................... 86
3.2 RESULTS....................................................................................................................... 88
3.2.1 Cardiolipin-LUVs accelerate IAPP misfolding ...................................................... 88
3.2.2 Physiological salt concentrations rectify dose-dependent acceleration of IAPP
misfolding .............................................................................................................................. 91
3.2.3 IAPP undergoes structural transitions through α-helical conformations upon
interaction with Cardiolipin-LUVs ........................................................................................ 93
3.2.4 Transmission Electron Microscopy of IAPP incubated with Cardiolipin-LUVs ... 94
3.3 DISCUSSION ................................................................................................................ 96
3.4 REFERENCES ............................................................................................................... 99
3.5 ACKNOWLEDGEMENTS ......................................................................................... 103
3.6 AUTHOR CONTRIBUTION STATEMENT ............................................................. 103
3.7 COMPETING FINANCIAL INTERESTS STATEMENT ......................................... 104
3.8 ABBREVIATIONS ...................................................................................................... 104
3.9 METHODS................................................................................................................... 104
6

Chapter 4 Chaperone-like Activity of Mitochondrially-Derived Peptides Captures Diabetic
Amyloid Seeds ............................................................................................................................ 108
ABSTRACT ............................................................................................................................ 109
4.1 INTRODUCTION ........................................................................................................ 110
4.2 RESULTS..................................................................................................................... 112
4.2.1 MDPs inhibit islet amyloid polypeptide misfolding. ............................................ 112
4.2.2 EPR Spectroscopy reveals MDPs prevent loss of monomeric IAPP without
interactions with free IAPP .................................................................................................. 115
4.2.3 CD spectroscopy and TEM indicates MDPs prevent the misfolding of IAPP ..... 118
4.2.4 MDPs, HNG and SHLP2, prevent propagation by IAPP seeds ............................ 123
4.2.5 MDPs bind directly to misfolded IAPP seeds....................................................... 124
4.3 DISCUSSION .............................................................................................................. 126
4.4 REFERENCES ............................................................................................................. 128
4.5 ACKNOWLEDGEMENTS ......................................................................................... 132
4.6 AUTHOR CONTRIBUTION STATEMENT ............................................................. 132
4.7 COMPETING FINANCIAL INTERESTS STATEMENT ......................................... 132
4.8 ABBREVIATIONS ...................................................................................................... 133
4.9 METHODS................................................................................................................... 133
Appendix A Lysine Acetylation is a Mechanism for Regulating the Interaction between Proteins
and Membranes ........................................................................................................................... 138
A.1 SUMMARY ..................................................................................................................... 139
A.2 RESULTS ......................................................................................................................... 141
A.2.1 Bioinformatics analysis reveals functional specificity for lysine acetylation at the
membrane binding interface of membrane binding domains .............................................. 141
A.2.2 Mimicking lysine acetylation in Amphiphysin inhibits membrane remodeling by
reducing membrane binding affinity and dissociates amphiphysin from tubular networks in
cells……………………………………………………………………………………….. 150
A.2.3 Mimicking lysine acetylation in Eps Homology Domain 2 prevents lipid remodeling,
reduces membrane binding affinity and leads to dissociation from the plasma membrane 153
7

A.2.4 Mimicking lysine acetylation in the C2A domain of Synaptotagmin 1 reduces
membrane binding affinity .................................................................................................. 155
A.2.5 Mimicking lysine acetylation of Amphiphysin disrupts the T-tubule network and
causes a flight defect in D. melanogaster ............................................................................ 156
A.3 DISCUSSION .................................................................................................................. 158
A.4 REFERENCES ................................................................................................................. 162
A.5 AUTHOR CONTRIBUTION STATEMENT.................................................................. 166
A.6 METHODS ....................................................................................................................... 167
CONCLUDING REMARKS ...................................................................................................... 175
REFERENCES ........................................................................................................................ 179

8

ABSTRACT
The misfolding and aggregation of proteins is associated with some of the most devastating
diseases of the modern era, including type 2 diabetes mellitus, Alzheimer disease and Parkinson
disease. In type 2 diabetes mellitus (T2DM), the misfolding of the 37-amino acid peptide, islet
amyloid polypeptide (IAPP), is associated with cytotoxicity of insulin producing β-cells. It is
thought that efforts to intervene in the misfolding process will lead to the development of
therapeutics for diseases such as T2DM. To facilitate these efforts, it is important to identify
conformations of proteins involved in disease-associated misfolding cascades. In this thesis we
investigate the misfolding of IAPP using a combination of thioflavin T fluorescence (ThT),
circular dichroism (CD), electron paramagnetic resonance spectroscopy (EPR), and transmission
electron microscopy (TEM). We find that liposomes loaded with anionic lipids and lipid-like
molecules associated with T2DM risk factors (e.g. obesity and plastics exposure), such as
phosphatidic acid (PA), oleic acid (OA), a fatty acid, and monobenzylester phthalate (MBzP), a
monophthalate ester, dramatically enhance the misfolding of IAPP. Such membrane interactions
with IAPP lead to the rapid formation of a membrane-bound and metastable α-helical intermediate
that slowly transitions into fibrillar amyloid on the surface of the membrane. Due to the
mitochondrial dysfunction seen in T2DM, we also investigated the interaction between the
negatively charged cardiolipin (CL), a uniquely mitochondrial lipid and IAPP. We found that CL
enhances the misfolding of IAPP in a manner similar to that of PA, OA and MBzP. These data
are consistent with the hypothesis that the ability to induce membrane-mediated misfolding of
IAPP is a generic trait of negatively charged membranes. This highlights the importance of the α-
helical conformation of IAPP as a drug development target since multiple risk factors for T2DM,
9

as well as mitochondria-like membranes, induce misfolding of IAPP through the same α-helical
conformer.
We then investigated the ability of two mitochondrial derived peptides (MDPs) to alter the course
of IAPP misfolding and found that both MDPs inhibit misfolding by direct, chaperone-like
interactions with misfolded amyloid seeds. This chaperone-like activity makes these MDPs
exciting new prospects for development as T2DM therapeutics since they provide the opportunity
to target key conformations along the misfolding pathway without necessitating atomistic level
characterization of the conformers themselves.
Taken together, this work focuses on two key targets along the IAPP misfolding pathway. First,
we highlight the importance of the α-helical intermediate of IAPP formed during membrane-
mediated misfolding. Second, we characterize two naturally occurring peptides of the MDP
family with chaperone-like activities that target misfolded IAPP conformations to prevent
propagation by misfolded templates. These studies are a part of larger, ongoing efforts to
understand the amyloid misfolding process and use that knowledge to develop therapeutics that
will one day lead to a cure for the millions of people suffering from T2DM and other protein
misfolding diseases.

10

Chapter 1
Introduction
1.1 Protein Misfolding in Human Disease
The three dimensional structures adopted by proteins are fundamental to the functions that they
perform. The ability to switch between various structural conformations is a feature that proteins
use to modify their function or performance. In some cases, however, proteins adopt dysfunctional
conformations that lead to aggregation of the protein. As a consequence, these proteins lose their
appropriate functions while gaining, in many cases, new functions that are toxic to the host cells
(1). When uncontrolled by cellular protein homeostasis machinery, these events can lead to cellular
and tissue level defects that manifest into diseases (2). These are generally referred to as protein-
misfolding diseases, some of which number among the most prevalent human diseases. All told,
approximately fifty human diseases are associated with the misfolding of what are normally
soluble and functional proteins or peptides (1). Type 2 diabetes mellitus (T2DM), Alzheimer
disease (AD) and Parkinson disease (PD) represent the most common protein-misfolding diseases
and account for more than 35 million cases in the US alone (3–5), more than 10% of the US
population.
Protein-misfolding diseases can be classed into three different categories based on their etiologies:
sporadic, hereditary, and transmissible. T2DM, AD and PD are predominantly sporadic in
occurrence, though in the cases of AD and PD there are examples of hereditary forms of the
diseases (Table 1.1) (6). Some amyloid diseases are exclusively hereditary. An example of which
is Huntington disease (HD), where CAG repeat expansion within the huntingtin gene results in a
11

huntingtin protein that misfolds, causing the death of neurons in the striatum. Some cases of
amyloid diseases occur following human-to-human transmission of the misfolded protein. The best
characterized cases of this involve the transmission of the the prion protein, PrP
Sc
, from an infected
to a naïve person, which causes the spongiform encephalopathies, Creutzfeldt-Jacob disease (CJD)
and kuru. This accounts for ~5% of spongiform encephalopathy cases, whereas sporadic and
hereditary case frequencies are 85% and 10%, respectively (6). Transmissibility from organism to
organism is not a trait commonly associated with T2DM, AD, PD, or HD. In the cases of AD and
PD, however, successful transmission of misfolded proteins to naïve animals has brought about
disease in the laboratory setting (7–10).
In most cases, the process of misfolding leads to a conversion of the protein in question into the
archetypal protein aggregate known as an amyloid fibril. Likewise, in most cases the development
of amyloid fibrils leads to cellular toxicity of a single or small group of cell type(s) (1). Yet the
relationship between fibrillization and toxicity is complex. A growing body of data suggests that
toxicity associated with protein misfolding is caused by intermediates developed during the
misfolding process and that fibrils themselves cause relatively little harm to cells (1,11–17). The
precise nature of the toxic intermediates as well as their mechanisms of toxicity remain an
important field of study. It is thought that therapies for protein-misfolding diseases will be found
by developing agents that prevent misfolding and thus, misfolding-associated toxicity. In fact,
while the general lack of therapeutics for protein-misfolding diseases remains a major shortcoming
of modern medicine, the drug Tafamidis [2-(3,5-dichloro-phenyl)-benzoxazole-6-carboxylic acid]
represents both a hope and proof-of-principle that therapeutics for amyloid diseases can be
successful. Tafamidis, which is currently used to treat familial amyloid polyneuropathy (FAP),
12

stabilizes transthyretin (TTR), preventing its misfolding and aggregation, ultimately stopping
disease progression in patients (18).
Table 1.1

Table 1.1 Common protein-misfolding diseases and their targets.

13

1.2 Amyloid Structure
Amyloid formation generally progresses from a monomeric starting point via what is thought to
be an array of multimeric intermediate states to the terminal fibrillar forms found within amyloid
deposits. The proteins found within amyloid deposits vary widely in primary structure between
diseases. In addition, these polypeptides also vary in their secondary structure while in their non-
amyloid, native conformations (1). Despite this heterogeneity of sequence and secondary structure,
intermediates that develop along the misfolding pathway share certain conformational traits.
Likewise, amyloid fibrils formed from different proteins also display multiple degrees of
conformational homogeneity. These points are highlighted by the work of Glabe and associates,
who developed a number of conformationally-specific antibodies that recognize intermediates
and/or fibrils from a variety of protein-misfolding diseases without showing significant affinity to
the monomeric proteins of which they are made (19). Furthermore, amyloid fibrils themselves
have stereotyped structural elements. In vitro studies of fibril morphology using transmission
electron microscopy and atomic force microscopy have revealed that fibrils are commonly
composed of 2 – 5 nm diameter protofilaments wound together to form 7 – 13 nm, non-branching
fibers (Figure 1.2.1) (20). X-ray diffraction, nuclear magnetic resonance spectroscopy and electron
paramagnetic resonance spectroscopy of amyloid fibrils have revealed that the protofilaments
themselves are arranged such that β-strands from individual units are oriented perpendicular to the
long axis of the fibril in what is described as a ‘cross-β’ pattern, often in a parallel, in-register
fashion (21–26). In much the same way that conformationally-specific antibodies recognize
epitopes of amyloid misfolding intermediates or fibrils across species, the amyloid fibril-specific
dyes, thioflavin T and Congo Red, bind amyloid fibrils from all species of amyloid. Thus, while
14

the native forms of amyloidogenic proteins are distinct entities, the misfolded structures they adopt
are relatively generic, especially at the level of the fibril (Figure 1.2.2).
Figure 1.2.1

Figure 1.2.1 – Islet amyloid polypeptide fibril structure modelled from electron paramagnetic
resonance spectroscopy data. A, B. Orthogonal views of an individual IAPP unit as it would be
structured within the amyloid fibril as viewed along the fibril axis (a) and rotated 90 ° (b) to
emphasize β-strand staggering. C, D. Orthogonal views of an IAPP protofilament model showing
the staggered orientation of β-strands. E. Model of IAPP protofilament containing 101 peptides
15

demonstrates a left handed helix with an ~90° turn between the red boxes. Figure adapted from
(27).
Figure 1.2.2

Figure 1.2.2 – Transmission electron microscopy of amyloid fibrils from four distinct protein-
misfolding diseases. A. Islet amyloid polypeptide fibril; type 2 diabetes mellitus. Image adapted
from (27). B. α-Synuclein fibril; Parkinson disease. C. Amyloid-β fibril; Alzheimer disease. D.
Huntingtin exon 1 (polyglutamine expansion of 46); Huntington disease. Scale bars = 200 nm.

16

1.3 Protein-Misfolding Kinetics
The processes that underlie the misfolding and aggregation of amyloids are complex. Accordingly,
a detailed mechanistic understanding of the steps that lead to misfolding is crucial to the
development of successful strategies to inhibit amyloid misfolding. The use of kinetics studies to
gather and formalize quantitative information about the misfolding process has helped to elucidate
the mechanistic underpinnings of protein aggregation (Figure 1.3.1). Methods such as time-
resolved fluorescence spectroscopy of fibril formation in the presence of thioflavin dyes have
helped to reveal that amyloid formation occurs in a stereotyped sigmoidal reaction time course
consisting of three main phases: a lag phase, growth phase and plateau phase (28,29). The presence
of a lag phase found in plots of reaction kinetics indicates that amyloid formation features a
nucleation step akin to that of crystallization. Nucleation events ultimately lead to the formation
of multimeric entities with prototypical amyloid cross-β structural features, in essence nascent
fibrils (30), however, whether the first steps of nucleation involve prototypical amyloid
conformations is not clear. The hydrophobic regions of the terminal peptides at the ends of the
nascent fibril become exposed to the aqueous environment, providing a thermodynamic driving
force that spurs addition of new monomeric units to the exposed fibril ends from the available pool
of soluble monomer (31). Secondary processes such as fragmentation of fibrils and surface
catalyzed secondary nucleation contribute to shortening of the lag phase and acceleration of the
growth phase (Figure 1.3.1). During the growth phase the free pool of naïve amyloid is consumed
rapidly. As the available pool of free protein dwindles, the growth rate tapers off and the final,
plateau phase begins, which represents an equilibrium state (31).
The use of formalized in vitro chemical kinetics studies to define the reaction mechanisms
involved in amyloid misfolding gives important information that can be applied to predicting how
17

these proteins will behave in in vivo environments. In terms of the study of potential therapeutics,
the ability to delay misfolding of an amyloid protein in an in vitro experiment where concentrations
of protein are extraordinarily high and time courses short relative to what may be encountered in
a living organism is an important indicator of the therapeutic agent’s potential likelihood of
success. Indeed, the time course of an in vitro experiment is substantially shorter than the decades
it takes for diseases like T2DM or AD to manifest, in part due to the use of high protein
concentrations.
Figure 1.3.1

Figure 1.3.1 – Formalized kinetics approaches to describing amyloid formation. A. Schematic of
the microscopic mechanisms of primary and secondary nucleation of amyloid fibrilization.
Monomers are shown in green, fibrils are shown in blue. Fibril elongation occurs along the fibril
18

axis, denoted by blue dots. Nucleation events are highlighted in yellow. B. Prototypical sigmoidal
reaction time course of an amyloid fibrilization reaction with three phases, a lag phase that is
described by τlag, a growth phase described by rmax, and a plateau phase. C. Proposed integrated
rate laws used for describing relative contributions of nucleating mechanisms at work during a
fibrilization reaction. Figure adapted from (31).

19

1.4 Islet Amyloid Misfolding as a Pathogenic Component in Type 2 Diabetes Mellitus
Type 2 diabetes mellitus is the most prevalent of the ~50 human protein misfolding diseases.
Worldwide, 422 million people were estimated to have diabetes in 2014 (32). In the United States
alone, 29.1 million Americans (9.3% of the U.S. population) were estimated to have diabetes in
2014, of which T2DM accounted for 90 – 95% (33). Diabetes remained the 7
th
leading cause of
death in the United States in 2010 and is also a significant contributor to cardiovascular disease-
related deaths including heart attacks and strokes. Other complications of diabetes include
blindness, kidney failure and non-traumatic lower-limb amputation (33). With an estimated 86
million Americans having prediabetes and 1.4 million new diabetes diagnoses per year (33),
T2DM represents a growing burden to society and a massive drain on healthcare resources in the
United States. The lack of therapeutic, disease-modifying treatments for T2DM only underscores
the need to better understand the disease process and to make extensive efforts to turn this
understanding into new treatment modalities.
The longstanding view of T2DM as a disease primarily of insulin resistance has given way in
recent years to the understanding that the essential dysfunction lies with an abnormal response by
the insulin producing β-cell to increased insulin demand (34). Whereas normal endocrine gland
responses to increased demand is hyperplastic growth and increased secretion to meet needs, in
diabetic patients this response is blunted. The blunted response of the β-cell in T2DM is in part
due to loss of responsiveness to energetic stimuli and in part due to a deficit in the absolute number
of β-cells (34). Clinical manifestations such as impaired fasting glucose levels are reported in
individuals where 50% or greater of the β-cell mass has been lost (35). In pancreata from human
diabetic patients, the loss of β-cell mass is concomitant with the deposition of hyaline deposits
composed primarily of amyloid, a finding that has been recognized for over 100 years (36). The
20

identity of the material within the deposits, however, remained a mystery until 1987 when two
independent groups reported the presence of islet amyloid polypeptide (IAPP) (37,38) within the
amyloid deposits.
It is now well-described that IAPP, a 37 amino acid peptide normally packaged within insulin
secretory granules, is the primary constituent of the amyloid deposits and is thought to play a
critical role in the pathogenesis of T2DM (39). As amyloid deposits are found in a majority of
diabetic patients (up to 90%), but also found in some individuals without clinical diabetes, a causal
role for IAPP in islet dysfunction was initially discounted (40,41). A number of lines of evidence
(detailed below) have helped to clarify the apparent discrepancy and place IAPP, and in particular
IAPP misfolding, at the etiological center of the β-cell defect.
Microscopic studies of diabetic islets show that a correlation exists between loss of β-cell mass
and deposition of islet amyloid. In general, those islets with IAPP deposition lack β-cell mass,
whereas those without deposition appear relatively normal (42,43). IAPP itself is cytotoxic since
both treatment with and expression of IAPP induces cell death via apoptotic changes in β-cells
(13,35,44). Interspecies comparisons among IAPP orthologs has implicated the process of IAPP
misfolding as critical for this cytotoxic function (13,40). Primate IAPP, including human IAPP,
has an intrinsic propensity to misfold that is not shared by rodent forms of IAPP such as rat and
mouse. This is generally attributed to the presence of proline residues in the rodent forms of IAPP
not present in primate IAPP that restrict flexibility in such a way as to prevent oligomerization and
subsequent fibrilization. This loss of misfolding phenotype is correlated with a loss of cytotoxic
phenotype. In addition, rats and mice do not spontaneously develop T2DM.
21

Reports from transgenic mouse and rat models expressing human IAPP have further revealed that
expression of human IAPP predisposes mice and rats to developing spontaneous diabetes (45–50).
Frequency of the manifestation of diabetic phenotypes is increased by induction of insulin
resistance via pharmacological means or obesity (46,49,51). Homozygosity in mice expressing
human IAPP also substantially increases spontaneous diabetes frequency (45), which implies that
the gene dosage of human IAPP positively correlates with disease occurrence. Rodents transgenic
for human IAPP also exhibit a deficit in β-cell mass secondary to increased β-cell apoptosis,
hyperglycemia, and impaired insulin secretion among other T2DM traits (52). Interestingly, as in
the human population, not all diabetic animals have demonstrable amyloid deposits. Instead,
positive staining with antibodies that target oligomeric intermediates of amyloid conformers
appear to be a more reliable indicator of amyloid toxicity (13). These experiments not only help to
clarify the discrepancy between clinical diabetes and the presence or absence of islet amyloid, but
also reveal that intermediates of misfolding (and thus the misfolding process itself) are sufficient
and likely primarily responsible for β-cell toxicity and the diabetic phenotype. A similar correlate
in humans was discovered in Japan where individuals carrying the S20G mutation in the IAPP
gene were found to be predisposed to early onset T2DM and reported strong family histories of
diabetes (53). Studies into the effect of the S20G mutation on IAPP revealed that S20G IAPP
displays an increased propensity towards misfolding and cytotoxicity (53–55). Together these data
argue that the presence of misfolding-competent IAPP is both necessary and sufficient for the
development of spontaneous type 2 diabetes mellitus. They further put to rest early controversy
regarding the role of IAPP in T2DM disease pathogenesis.
22

Figure 1.4.1

Figure 1.4.1 – IAPP ortholog alignment. Primary sequence alignments for human, S20G mutant,
monkey, cat, dog, mouse and rat IAPP. Conserved residues are represented with a dot. Red
lettering highlights the amyloidogenic region of IAPP where proline mutations in rodent IAPP
render the peptide unable to fibrilize. Figure adapted from (13).
23

Figure 1.4.2

Figure 1.4.2 – Amyloid staining in islets reveals β-cell mass is replaced by amyloid. A, B. Islet
amyloid deposition from a human type 2 diabetic stained with (a) Congo Red and viewed by
partially cross-polarized light shows apple green birefringence, or (b) with Thioflavin S and shows
amyloid (green) replacing insulin (red) secreting β-cells. C. Islet from a non-diabetic human
stained with Thioflavin S shows little amyloid. D. Islet taken from a human IAPP transgenic mouse
on a high fat diet shows replacement of insulin secreting β-cells with Thioflavin S positive islet
amyloid. E. Islet taken from a nontransgenic mouse on the same diet as in (D) shows no islet
amyloid. Figure adapted from (39).

24

1.5 Protein Misfolding Pathways of Islet Amyloid Polypeptide: Membrane-Mediated
Misfolding
In solution, islet amyloid polypeptide misfolds in a manner akin to other amyloids in many ways,
but differs in certain significant aspects. IAPP aggregation proceeds along a pathway wherein
monomeric, soluble IAPP misfolds into oligomeric intermediates and ultimately into insoluble
amyloid fibrils. The first step is slow and may sample a micelle structure (56) that allows
nucleating interactions to occur at the center of the peptide (residues 12 – 21) (Figure 1.5.1) (57).
Unlike other well-studied amyloids (58), IAPP does not readily form metastable oligomers, instead
forming transient intermediates that have proven exceedingly difficult to isolate and stabilize
(59,60). Thus, thermodynamically speaking, the misfolding of IAPP in solution is a process that
flows downhill to the fibrillar form, while sampling without significantly accumulating misfolding
intermediates.
The effects of membrane interactions with human IAPP profoundly change the nature of IAPP
misfolding. The duration of the lag phase of IAPP fibrilization, as with other amyloids, is primarily
dependent upon the rate of spontaneous nucleation. Studies performed with IAPP in the presence
of liposomes containing phosphatidylserine or phosphatidylglycerol have demonstrated that the
presence of these lipids dramatically reduces the span of the lag phase, thus accelerating IAPP
misfolding (61,62). In addition, IAPP gains significant helicity in the region of amino acids 9 to
22 in the presence of phosphatidylserine containing membranes (61,63) (Figure 1.5.2a). It is
thought that the membrane-mediated misfolding of IAPP accelerates misfolding because of a high
frequency of membrane-driven nucleation events (Figure 1.5.2b). These nucleation events are
thought to increase for two reasons: 1) coulombic interactions between the positively charged
residues in IAPP and the negatively charged lipids pull IAPP into the membrane. As a
25

consequence, the local membrane concentration of IAPP increases. Furthermore, cationic residues
on IAPP that are mutually repulsive become bound by the lipids, reducing repulsion between IAPP
molecules permitting greater frequency of IAPP-IAPP interactions. The reduction in
dimensionality as IAPP moves from a three dimensional diffusion volume to a two dimensional
diffusion plane further serves to concentrate IAPP and promote interactions. 2) The helix of IAPP
extends from amino acids 9 to 22, but the highly amyloidogenic region beyond 22 remains
unstructured and exposed in an environment that lends itself to promoting the formation of
misfolded β-sheet structures (61,63).
Studies of IAPP with membranes and in cellular environments suggest that disruptions of
membrane integrity are responsible for IAPP’s observed cytotoxicity (64–69). These studies and
others (70) support two leading hypotheses regarding the mechanism of membrane perturbation,
the pore formation hypothesis and the membrane fragmentation/detergent hypothesis (Figure
1.5.3). In the first, an amyloid forms a transmembrane channel whereas in the second, amyloids in
the membrane act as detergents pulling lipid from the membrane. In both cases, the perturbation
allows inappropriate and uncontrolled ion flux that leads to membrane and cellular dysfunction. In
light of these findings, it appears that interactions between human IAPP and membrane lipids are
mutually destructive.

26

Figure 1.5.1

Figure 1.5.1 – Model of the mechanisms of IAPP misfolding in aqueous solution. IAPP, as an
intrinsically disordered peptide in solution (i) may sample a micellar state (ii). Misfolded
intermediates such as the helical intermediate (iii) and β-sheet oligomer (iv) are thought to exist,
however, isolation and characterization of these structures has not yet been accomplished and thus
their presence and correct order in this schematic remain uncertain. Formation of amyloid fibrils
(v) from a nucleus occurs rapidly in the growth phase. Figure adapted from (57).
27

Figure 1.5.2

Figure 1.5.2 – EPR-based structure of membrane-bound IAPP suggests a membrane-mediated
misfolding mechanism. A. Schematic of IAPP helix formation in an anionic (phosphatidylserine)
membrane. The core helical region extends from T9 to S20. The helix extends through N22,
however, increased mobility in the EPR spectra at 21 and 22 are consistent with a fraying of the
C-terminal region of the helix. Regions extending before and after the helix were not specifically
ordered. The depth of insertion into the bilayer was determined using EPR-based depth
measurements. Dark spheres represent phosphates. B. Cartoon model of membrane-mediated
misfolding of IAPP. Figure adapted from (63).

28

Figure 1.5.3

Figure 1.5.3 – Two leading hypotheses of IAPP-mediated membrane disruption. A. Pore-forming
hypothesis indicates that IAPP could form a structured pore in the membrane. Two proposed
mechanisms are schematically represented, the barrel-stave pore (left) and the toroidal pore (right).
B. The membrane fragmentation model or detergent model of IAPP-mediated membrane
disruption are shown. Here, after a threshold concentration of IAPP is reached in the membrane,
induced curvature in the membrane allows for gaps in the membrane to open and small spherical
structures to be removed from the membrane, leading eventually to disintegration of the bilayer.
Figure adapted from (70).
1.6 Approaches to Modifying Protein Misfolding Disease Progression
1.6.1 Targeting Protein Misfolding
One of the great challenges to modern medicine is the development of therapeutic approaches to
protein misfolding diseases. Of the ~50 known human protein-misfolding diseases, only one,
29

familial amyloidotic polyneuropathy (FAP), has seen the development of a pharmaceutical agent
that can stop disease progression (18,71). In that case, a detailed mechanistic understanding of the
misfolding pathway led to the development of a small molecule, Tafamidis, that stabilized the
protein transthyretin in a conformation that prevented its misfolding and aggregation. It is hoped
that studying and engineering inhibitors to the misfolding process will lead to the development of
pharmaceuticals that can stop disease progression for many other protein-misfolding diseases,
including T2DM.
Every step along the misfolding pathway provides potential opportunities to alter the misfolding
process. Directly binding monomeric amyloids such as IAPP could help to prevent or discourage
inappropriate interactions with other IAPP molecules. While in theory this could prove successful,
there are practical drawbacks to this approach. Most amyloid proteins are likely to perform normal
physiological functions in their native state and interfering with the monomer may alter the ability
of the protein to carry out that task. In addition, in order to effectively neutralize the monomer,
near stoichiometric equivalents of inhibitors are expected to be required which may necessitate
large quantities of the inhibitor be used.
Targeting the misfolded forms of the amyloid offers another, arguably more attractive solution.
Ideally, specifically targeting the primary nucleation process by capturing and perhaps refolding
the nucleated seeds into the naïve state, would allow the monomer population to continue in its
normal functional role. Similarly, later intermediate products of the misfolding reaction, such as
larger oligomeric and protofibrillar forms that are strongly associated with toxicity could be
captured and perhaps disaggregated to prevent the apoptotic process from occurring in vulnerable
tissues. These conformations have been difficult to specifically target as many of the intermediates
that ultimately cause the various forms of amyloid diseases lack defined structures that can be used
30

to design synthetic small molecules that would target them (57). The fiber is also a theoretical
target, but as it is generally thought to play little role in toxicity (13), it is not considered ideal. In
fact, inhibition of fibrilization, but not formation of toxic oligomers by rifampicin still leads to
IAPP toxicity of β-cells (72). In the case of Tafamidis, the Kelly group identified and targeted an
off-pathway, non-toxic conformation with success, a homo-tetrameric TTR (18). It is possible that
in other diseases, analogous structures exist that can be stabilized. In the case of IAPP, such an
analogous structure might be the membrane-bound helix (63).
There are three main approaches to the inhibition of protein misfolding that are currently under
investigation that are designed to target many of the above steps. They include the use of small
molecules, peptides (as well as peptidomimetics), and heat shock proteins and are discussed below.
1.6.2 Small Molecule Inhibitors of IAPP
Small molecules are often found in nature, the product of any number of organisms that can be
isolated and repurposed for various tasks, while their synthetic counterparts can be engineered for
specific purposes. A handful of naturally occurring small molecules have showed promise towards
inhibiting the misfolding of IAPP including epigallocatechin-gallate (EGCG), curcumin, and
resveratrol (57) (Figure 1.6).
EGCG is a flavonoid found in green tea under investigation as a candidate to prevent various
amyloid diseases, including AD, PD and HD. It is thought that EGCG directs IAPP into an off-
pathway aggregate, but also shows the capacity to destabilize preformed fibrils into smaller β-
sheeting containing fibers (57). Curcumin is a natural product found enriched in turmeric with
numerous beneficial properties ranging from antioxidant to anti-microbial and anti-amyloidogenic
activities. Curcumin has been found to inhibit the misfolding of IAPP and protect from toxicity
31

(73). It shows little effect on preformed fibrils, but is instead thought to destabilize helical
intermediates (74). Resveratrol, found in grapes and red wine, has also been found to have
beneficial effects on the misfolding and toxicity of IAPP. It reduces fibrilization of IAPP, instead
leading to small, non-toxic and off-pathway oligomeric structures, even showing effects in the
presence of anionic membranes (75). Other naturally occurring small molecules are under
investigation as well, some of which may be useful in preventing T2DM disease progression
(57,76,77).
Figure 1.6

Figure 1.6 – Structures of some of the best characterized small molecule IAPP misfolding
inhibitors. Figure adapted from (57).
1.6.3 Proteins and Peptides as inhibitors of IAPP misfolding
Proteins and peptides represent another class of molecules being investigated for use in preventing
toxicity associated with IAPP misfolding, however, none have yet advanced to clinical trials. One
32

such approach has used conformationally-specific antibodies (often developed against other
misfolded amyloids) to target IAPP misfolding (19,78). This has met with some success in
protecting cells from IAPP toxicity. A more common approach has been to design peptide-based
inhibitors with identical or slightly modified sequences taken from amyloidogenic regions of IAPP
(79–82). Most have focused on the NFGAILSS region of IAPP and have also met with some
success in protecting cells from IAPP toxicity. A hybrid approach has also emerged wherein
amyloidogenic sequences of IAPP were grafted into antigen-recognition domains of antibodies
(83). While these approaches have shown that cells can be protected from extracellular IAPP, they
intrinsically suffer from difficulties delivering peptides or antibodies into cells. Thus, while they
may have the capacity to interact with extracellular amyloid, toxicity associated with intracellular
amyloid presents a significant barrier to translation of these efforts into successful therapeutics.
1.6.4 Chaperones as Inhibitors of IAPP Misfolding
More recent efforts geared towards targeting the misfolding pathway of IAPP have taken
advantage of naturally occurring protein chaperones or heat shock proteins. These proteins
function endogenously to combat the natural predisposition of proteins to adopt undesirable folds
that lead to protein aggregation. Relatively low concentrations of HSPs and chaperones can inhibit
the misfolding of high concentrations of IAPP by specifically targeting misfolded IAPP
subpopulations (84,85). As a consequence, the bulk of the IAPP remains in its natively unfolded
conformation(s), protected from the influence of misfolded species. These and other studies have
demonstrated protection of β-cells against internal expression and external application of IAPP
(84–88). This field of research represents a new and exciting approach to the treatment of T2DM.
The use of proteins that do not bind monomeric IAPP allow for high potency to be achieved as
reactions with the bulk of the naïve IAPP are avoided in favor of interactions with a smaller
33

population of misfolded species. An additional benefit of this approach is that many of these
inhibitors are found naturally expressed in or secreted by cells. This allows for additional flexibility
in the approaches to treatment with these agents. Beyond direct delivery of the chaperone into the
blood stream, methods to induce intracellular expression may be developed.
1.7 REFERENCES
1. Chiti F, Dobson CM. Protein Misfolding, Functional Amyloid, and Human Disease. Annu
Rev Biochem. 2006;75(1):333–66.
2. Hipp MS, Park S-H, Hartl FU. Proteostasis impairment in protein-misfolding and -
aggregation diseases. Trends Cell Biol. 2014 Sep;24(9):506–14.
3. Alex ADA 1701 NBS, ria, 1-800-Diabetes V 22311. Statistics About Diabetes [Internet].
American Diabetes Association. [cited 2016 May 28]. Available from:
http://www.diabetes.org/diabetes-basics/statistics/
4. Latest Alzheimer’s Facts and Figures [Internet]. Latest Facts & Figures Report |
Alzheimer’s Association. 2013 [cited 2016 May 28]. Available from:
http://www.alz.org/facts/overview.asp
5. Statistics on Parkinson’s - Parkinson’s Disease Foundation (PDF) [Internet]. [cited 2016
May 28]. Available from: http://www.pdf.org/en/parkinson_statistics
6. Prusiner SB. Biology and genetics of prions causing neurodegeneration. Annu Rev Genet.
2013;47:601–23.
34

7. Meyer-Luehmann M, Coomaraswamy J, Bolmont T, Kaeser S, Schaefer C, Kilger E, et al.
Exogenous induction of cerebral beta-amyloidogenesis is governed by agent and host.
Science. 2006 Sep 22;313(5794):1781–4.
8. Eisele YS, Obermüller U, Heilbronner G, Baumann F, Kaeser SA, Wolburg H, et al.
Peripherally applied Abeta-containing inoculates induce cerebral beta-amyloidosis.
Science. 2010 Nov 12;330(6006):980–2.
9. Luk KC, Kehm V, Carroll J, Zhang B, O’Brien P, Trojanowski JQ, et al. Pathological α-
synuclein transmission initiates Parkinson-like neurodegeneration in nontransgenic mice.
Science. 2012 Nov 16;338(6109):949–53.
10. Masuda-Suzukake M, Nonaka T, Hosokawa M, Oikawa T, Arai T, Akiyama H, et al. Prion-
like spreading of pathological α-synuclein in brain. Brain J Neurol. 2013 Apr;136(Pt
4):1128–38.
11. Walsh DM, Klyubin I, Fadeeva JV, Cullen WK, Anwyl R, Wolfe MS, et al. Naturally
secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term
potentiation in vivo. Nature. 2002 Apr 4;416(6880):535–9.
12. Koffie RM, Meyer-Luehmann M, Hashimoto T, Adams KW, Mielke ML, Garcia-Alloza
M, et al. Oligomeric amyloid beta associates with postsynaptic densities and correlates with
excitatory synapse loss near senile plaques. Proc Natl Acad Sci U S A. 2009 Mar
10;106(10):4012–7.
13. Haataja L, Gurlo T, Huang CJ, Butler PC. Islet amyloid in type 2 diabetes, and the toxic
oligomer hypothesis. Endocr Rev. 2008 May;29(3):303–16.
35

14. Cremades N, Cohen SIA, Deas E, Abramov AY, Chen AY, Orte A, et al. Direct observation
of the interconversion of normal and toxic forms of α-synuclein. Cell. 2012 May
25;149(5):1048–59.
15. Winner B, Jappelli R, Maji SK, Desplats PA, Boyer L, Aigner S, et al. In vivo demonstration
that alpha-synuclein oligomers are toxic. Proc Natl Acad Sci U S A. 2011 Mar
8;108(10):4194–9.
16. Eisenberg D, Jucker M. The amyloid state of proteins in human diseases. Cell. 2012 Mar
16;148(6):1188–203.
17. Bucciantini M, Giannoni E, Chiti F, Baroni F, Formigli L, Zurdo J, et al. Inherent toxicity
of aggregates implies a common mechanism for protein misfolding diseases. Nature. 2002
Apr 4;416(6880):507–11.
18. Johnson SM, Connelly S, Fearns C, Powers ET, Kelly JW. The transthyretin amyloidoses:
from delineating the molecular mechanism of aggregation linked to pathology to a
regulatory-agency-approved drug. J Mol Biol. 2012 Aug 10;421(2–3):185–203.
19. Kayed R, Head E, Thompson JL, McIntire TM, Milton SC, Cotman CW, et al. Common
structure of soluble amyloid oligomers implies common mechanism of pathogenesis.
Science. 2003 Apr 18;300(5618):486–9.
20. Serpell LC, Sunde M, Benson MD, Tennent GA, Pepys MB, Fraser PE. The protofilament
substructure of amyloid fibrils. J Mol Biol. 2000 Jul 28;300(5):1033–9.
36

21. Sunde M, Blake C. The structure of amyloid fibrils by electron microscopy and X-ray
diffraction. Adv Protein Chem. 1997;50:123–59.
22. Nilsson MR. Techniques to study amyloid fibril formation in vitro. Methods San Diego
Calif. 2004 Sep;34(1):151–60.
23. Margittai M, Langen R. Fibrils with parallel in-register structure constitute a major class of
amyloid fibrils: molecular insights from electron paramagnetic resonance spectroscopy. Q
Rev Biophys. 2008 Nov;41(3–4):265–97.
24. Wasmer C, Lange A, Van Melckebeke H, Siemer AB, Riek R, Meier BH. Amyloid fibrils
of the HET-s(218-289) prion form a beta solenoid with a triangular hydrophobic core.
Science. 2008 Mar 14;319(5869):1523–6.
25. Petkova AT, Ishii Y, Balbach JJ, Antzutkin ON, Leapman RD, Delaglio F, et al. A structural
model for Alzheimer’s beta -amyloid fibrils based on experimental constraints from solid
state NMR. Proc Natl Acad Sci U S A. 2002 Dec 24;99(26):16742–7.
26. Fitzpatrick AWP, Debelouchina GT, Bayro MJ, Clare DK, Caporini MA, Bajaj VS, et al.
Atomic structure and hierarchical assembly of a cross-β amyloid fibril. Proc Natl Acad Sci
U S A. 2013 Apr 2;110(14):5468–73.
27. Bedrood S, Li Y, Isas JM, Hegde BG, Baxa U, Haworth IS, et al. Fibril Structure of Human
Islet Amyloid Polypeptide. J Biol Chem. 2012 Feb 17;287(8):5235–41.
37

28. Nielsen L, Khurana R, Coats A, Frokjaer S, Brange J, Vyas S, et al. Effect of environmental
factors on the kinetics of insulin fibril formation: elucidation of the molecular mechanism.
Biochemistry (Mosc). 2001 May 22;40(20):6036–46.
29. Knowles TPJ, Waudby CA, Devlin GL, Cohen SIA, Aguzzi A, Vendruscolo M, et al. An
analytical solution to the kinetics of breakable filament assembly. Science. 2009 Dec
11;326(5959):1533–7.
30. Serio TR, Cashikar AG, Kowal AS, Sawicki GJ, Moslehi JJ, Serpell L, et al. Nucleated
conformational conversion and the replication of conformational information by a prion
determinant. Science. 2000 Aug 25;289(5483):1317–21.
31. Knowles TPJ, Vendruscolo M, Dobson CM. The amyloid state and its association with
protein misfolding diseases. Nat Rev Mol Cell Biol. 2014 Jun;15(6):384–96.
32. WHO | Diabetes [Internet]. WHO. [cited 2016 May 28]. Available from:
http://www.who.int/mediacentre/factsheets/fs312/en/
33. 2014 Statistics Report | Data & Statistics | Diabetes | CDC [Internet]. [cited 2016 May 28].
Available from: http://www.cdc.gov/diabetes/data/statistics/2014statisticsreport.html
34. Alejandro EU, Gregg B, Blandino-Rosano M, Cras-Méneur C, Bernal-Mizrachi E. Natural
history of β-cell adaptation and failure in type 2 diabetes. Mol Aspects Med. 2015
Apr;42:19–41.
38

35. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC. Beta-cell deficit and
increased beta-cell apoptosis in humans with type 2 diabetes. Diabetes. 2003 Jan;52(1):102–
10.
36. Opie EL. THE RELATION OE DIABETES MELLITUS TO LESIONS OF THE
PANCREAS. HYALINE DEGENERATION OF THE ISLANDS OE LANGERHANS. J
Exp Med. 1901 Mar 25;5(5):527–40.
37. Westermark P, Wernstedt C, O’Brien TD, Hayden DW, Johnson KH. Islet amyloid in type
2 human diabetes mellitus and adult diabetic cats contains a novel putative polypeptide
hormone. Am J Pathol. 1987 Jun;127(3):414–7.
38. Cooper GJ, Willis AC, Clark A, Turner RC, Sim RB, Reid KB. Purification and
characterization of a peptide from amyloid-rich pancreases of type 2 diabetic patients. Proc
Natl Acad Sci U S A. 1987 Dec;84(23):8628–32.
39. Hull RL, Westermark GT, Westermark P, Kahn SE. Islet amyloid: a critical entity in the
pathogenesis of type 2 diabetes. J Clin Endocrinol Metab. 2004 Aug;89(8):3629–43.
40. Westermark P. Amyloid in the islets of Langerhans: Thoughts and some historical aspects.
Ups J Med Sci. 2011 May;116(2):81–9.
41. Cho J, Kim J, Shin J, Shin J, Yoon K. β‐cell mass in people with type 2 diabetes. J Diabetes
Investig. 2011 Jan 24;2(1):6–17.
42. Westermark P, Wilander E. The influence of amyloid deposits on the islet volume in
maturity onset diabetes mellitus. Diabetologia. 1978 Nov;15(5):417–21.
39

43. Clark A, Wells CA, Buley ID, Cruickshank JK, Vanhegan RI, Matthews DR, et al. Islet
amyloid, increased A-cells, reduced B-cells and exocrine fibrosis: quantitative changes in
the pancreas in type 2 diabetes. Diabetes Res Edinb Scotl. 1988 Dec;9(4):151–9.
44. Tucker HM, Rydel RE, Wright S, Estus S. Human amylin induces “apoptotic” pattern of
gene expression concomitant with cortical neuronal apoptosis. J Neurochem. 1998
Aug;71(2):506–16.
45. Janson J, Soeller WC, Roche PC, Nelson RT, Torchia AJ, Kreutter DK, et al. Spontaneous
diabetes mellitus in transgenic mice expressing human islet amyloid polypeptide. Proc Natl
Acad Sci U S A. 1996 Jul 9;93(14):7283–8.
46. Soeller WC, Janson J, Hart SE, Parker JC, Carty MD, Stevenson RW, et al. Islet amyloid-
associated diabetes in obese A(vy)/a mice expressing human islet amyloid polypeptide.
Diabetes. 1998 May;47(5):743–50.
47. de Koning EJ, Morris ER, Hofhuis FM, Posthuma G, Höppener JW, Morris JF, et al. Intra-
and extracellular amyloid fibrils are formed in cultured pancreatic islets of transgenic mice
expressing human islet amyloid polypeptide. Proc Natl Acad Sci U S A. 1994 Aug
30;91(18):8467–71.
48. Yagui K, Yamaguchi T, Kanatsuka A, Shimada F, Huang CI, Tokuyama Y, et al. Formation
of islet amyloid fibrils in beta-secretory granules of transgenic mice expressing human islet
amyloid polypeptide/amylin. Eur J Endocrinol Eur Fed Endocr Soc. 1995 Apr;132(4):487–
96.
40

49. Höppener JW, Oosterwijk C, Nieuwenhuis MG, Posthuma G, Thijssen JH, Vroom TM, et
al. Extensive islet amyloid formation is induced by development of Type II diabetes mellitus
and contributes to its progression: pathogenesis of diabetes in a mouse model. Diabetologia.
1999 Apr;42(4):427–34.
50. Butler AE, Jang J, Gurlo T, Carty MD, Soeller WC, Butler PC. Diabetes due to a progressive
defect in beta-cell mass in rats transgenic for human islet amyloid polypeptide (HIP Rat): a
new model for type 2 diabetes. Diabetes. 2004 Jun;53(6):1509–16.
51. Couce M, Kane LA, O’Brien TD, Charlesworth J, Soeller W, McNeish J, et al. Treatment
with growth hormone and dexamethasone in mice transgenic for human islet amyloid
polypeptide causes islet amyloidosis and beta-cell dysfunction. Diabetes. 1996
Aug;45(8):1094–101.
52. Matveyenko AV, Butler PC. Beta-cell deficit due to increased apoptosis in the human islet
amyloid polypeptide transgenic (HIP) rat recapitulates the metabolic defects present in type
2 diabetes. Diabetes. 2006 Jul;55(7):2106–14.
53. Sakagashira S, Sanke T, Hanabusa T, Shimomura H, Ohagi S, Kumagaye KY, et al.
Missense mutation of amylin gene (S20G) in Japanese NIDDM patients. Diabetes. 1996
Sep;45(9):1279–81.
54. Sakagashira S, Hiddinga HJ, Tateishi K, Sanke T, Hanabusa T, Nanjo K, et al. S20G mutant
amylin exhibits increased in vitro amyloidogenicity and increased intracellular cytotoxicity
compared to wild-type amylin. Am J Pathol. 2000 Dec;157(6):2101–9.
41

55. Ma Z, Westermark GT, Sakagashira S, Sanke T, Gustavsson A, Sakamoto H, et al.
Enhanced in vitro production of amyloid-like fibrils from mutant (S20G) islet amyloid
polypeptide. Amyloid Int J Exp Clin Investig Off J Int Soc Amyloidosis. 2001
Dec;8(4):242–9.
56. Brender JR, Krishnamoorthy J, Sciacca MFM, Vivekanandan S, D’Urso L, Chen J, et al.
Probing the sources of the apparent irreproducibility of amyloid formation: drastic changes
in kinetics and a switch in mechanism due to micellelike oligomer formation at critical
concentrations of IAPP. J Phys Chem B. 2015 Feb 19;119(7):2886–96.
57. Pithadia A, Brender JR, Fierke CA, Ramamoorthy A. Inhibition of IAPP Aggregation and
Toxicity by Natural Products and Derivatives. J Diabetes Res [Internet]. 2016 [cited 2016
May 1];2016. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4662995/
58. Sengupta U, Nilson AN, Kayed R. The Role of Amyloid-β Oligomers in Toxicity,
Propagation, and Immunotherapy. EBioMedicine. 2016 Apr;6:42–9.
59. Nagel-Steger L, Owen MC, Strodel B. An Account of Amyloid Oligomers: Facts and
Figures Obtained from Experiments and Simulations. ChemBioChem. 2016 Apr
15;17(8):657–76.
60. Suzuki Y, Brender JR, Hartman K, Ramamoorthy A, G. Marsh EN. Alternative Pathways
of Human Islet Amyloid Polypeptide Aggregation Distinguished by 19F NMR-Detected
Kinetics of Monomer Consumption. Biochemistry (Mosc). 2012 Oct 16;51(41):8154–62.
61. Jayasinghe SA, Langen R. Lipid membranes modulate the structure of islet amyloid
polypeptide. Biochemistry (Mosc). 2005 Sep 13;44(36):12113–9.
42

62. Knight JD, Miranker AD. Phospholipid catalysis of diabetic amyloid assembly. J Mol Biol.
2004 Aug 27;341(5):1175–87.
63. Apostolidou M, Jayasinghe SA, Langen R. Structure of alpha-helical membrane-bound
human islet amyloid polypeptide and its implications for membrane-mediated misfolding. J
Biol Chem. 2008 Jun 20;283(25):17205–10.
64. Lashuel HA. Membrane permeabilization: a common mechanism in protein-misfolding
diseases. Sci Aging Knowl Environ SAGE KE. 2005 Sep 21;2005(38):pe28.
65. Janson J, Ashley RH, Harrison D, McIntyre S, Butler PC. The mechanism of islet amyloid
polypeptide toxicity is membrane disruption by intermediate-sized toxic amyloid particles.
Diabetes. 1999 Mar;48(3):491–8.
66. Mirzabekov TA, Lin MC, Kagan BL. Pore formation by the cytotoxic islet amyloid peptide
amylin. J Biol Chem. 1996 Jan 26;271(4):1988–92.
67. Green JD, Kreplak L, Goldsbury C, Li Blatter X, Stolz M, Cooper GS, et al. Atomic force
microscopy reveals defects within mica supported lipid bilayers induced by the
amyloidogenic human amylin peptide. J Mol Biol. 2004 Sep 17;342(3):877–87.
68. Engel MFM, Khemtémourian L, Kleijer CC, Meeldijk HJD, Jacobs J, Verkleij AJ, et al.
Membrane damage by human islet amyloid polypeptide through fibril growth at the
membrane. Proc Natl Acad Sci U S A. 2008 Apr 22;105(16):6033–8.
69. Brender JR, Dürr UHN, Heyl D, Budarapu MB, Ramamoorthy A. Membrane Fragmentation
by an Amyloidogenic Fragment of Human Islet Amyloid Polypeptide Detected by Solid-
43

State NMR Spectroscopy of Membrane Nanotubes. Biochim Biophys Acta. 2007
Sep;1768(9):2026–9.
70. Butterfield SM, Lashuel HA. Amyloidogenic protein-membrane interactions: mechanistic
insight from model systems. Angew Chem Int Ed Engl. 2010 Aug 2;49(33):5628–54.
71. Coelho T, Maia LF, Silva AM da, Cruz MW, Planté-Bordeneuve V, Lozeron P, et al.
Tafamidis for transthyretin familial amyloid polyneuropathy A randomized, controlled trial.
Neurology. 2012 Aug 21;79(8):785–92.
72. Meier JJ, Kayed R, Lin C-Y, Gurlo T, Haataja L, Jayasinghe S, et al. Inhibition of human
IAPP fibril formation does not prevent beta-cell death: evidence for distinct actions of
oligomers and fibrils of human IAPP. Am J Physiol Endocrinol Metab. 2006
Dec;291(6):E1317-1324.
73. Daval M, Bedrood S, Gurlo T, Huang C-J, Costes S, Butler PC, et al. The effect of curcumin
on human islet amyloid polypeptide misfolding and toxicity. Amyloid Int J Exp Clin
Investig Off J Int Soc Amyloidosis. 2010 Sep;17(3–4):118–28.
74. Sparks S, Liu G, Robbins KJ, Lazo ND. Curcumin modulates the self-assembly of the islet
amyloid polypeptide by disassembling α-helix. Biochem Biophys Res Commun. 2012 Jun
15;422(4):551–5.
75. Mishra R, Sellin D, Radovan D, Gohlke A, Winter R. Inhibiting islet amyloid polypeptide
fibril formation by the red wine compound resveratrol. Chembiochem Eur J Chem Biol.
2009 Feb 13;10(3):445–9.
44

76. Noor H, Cao P, Raleigh DP. Morin hydrate inhibits amyloid formation by islet amyloid
polypeptide and disaggregates amyloid fibers. Protein Sci Publ Protein Soc. 2012
Mar;21(3):373–82.
77. Li X-L, Xu G, Chen T, Wong Y-S, Zhao H-L, Fan R-R, et al. Phycocyanin protects INS-1E
pancreatic beta cells against human islet amyloid polypeptide-induced apoptosis through
attenuating oxidative stress and modulating JNK and p38 mitogen-activated protein kinase
pathways. Int J Biochem Cell Biol. 2009 Jul;41(7):1526–35.
78. O’Nuallain B, Hrncic R, Wall JS, Weiss DT, Solomon A. Diagnostic and therapeutic
potential of amyloid-reactive IgG antibodies contained in human sera. J Immunol Baltim
Md 1950. 2006 Jun 1;176(11):7071–8.
79. Porat Y, Mazor Y, Efrat S, Gazit E. Inhibition of islet amyloid polypeptide fibril formation:
a potential role for heteroaromatic interactions. Biochemistry (Mosc). 2004 Nov
16;43(45):14454–62.
80. Scrocchi LA, Chen Y, Waschuk S, Wang F, Cheung S, Darabie AA, et al. Design of peptide-
based inhibitors of human islet amyloid polypeptide fibrillogenesis. J Mol Biol. 2002 May
3;318(3):697–706.
81. Yan L-M, Tatarek-Nossol M, Velkova A, Kazantzis A, Kapurniotu A. Design of a mimic
of nonamyloidogenic and bioactive human islet amyloid polypeptide (IAPP) as nanomolar
affinity inhibitor of IAPP cytotoxic fibrillogenesis. Proc Natl Acad Sci U S A. 2006 Feb
14;103(7):2046–51.
45

82. Tatarek-Nossol M, Yan L-M, Schmauder A, Tenidis K, Westermark G, Kapurniotu A.
Inhibition of hIAPP amyloid-fibril formation and apoptotic cell death by a designed hIAPP
amyloid- core-containing hexapeptide. Chem Biol. 2005 Jul;12(7):797–809.
83. Ladiwala ARA, Bhattacharya M, Perchiacca JM, Cao P, Raleigh DP, Abedini A, et al.
Rational design of potent domain antibody inhibitors of amyloid fibril assembly. Proc Natl
Acad Sci U S A. 2012 Dec 4;109(49):19965–70.
84. Chien V, Aitken JF, Zhang S, Buchanan CM, Hickey A, Brittain T, et al. The chaperone
proteins HSP70, HSP40/DnaJ and GRP78/BiP suppress misfolding and formation of β-
sheet-containing aggregates by human amylin: a potential role for defective chaperone
biology in Type 2 diabetes. Biochem J. 2010 Nov 15;432(1):113–21.
85. Mannini B, Cascella R, Zampagni M, van Waarde-Verhagen M, Meehan S, Roodveldt C,
et al. Molecular mechanisms used by chaperones to reduce the toxicity of aberrant protein
oligomers. Proc Natl Acad Sci U S A. 2012 Jul 31;109(31):12479–84.
86. Peinado JR, Sami F, Rajpurohit N, Lindberg I. Blockade of islet amyloid polypeptide
fibrillation and cytotoxicity by the secretory chaperones 7B2 and proSAAS. FEBS Lett.
2013 Nov 1;587(21):3406–11.
87. Cadavez L, Montane J, Alcarraz-Vizán G, Visa M, Vidal-Fàbrega L, Servitja J-M, et al.
Chaperones ameliorate beta cell dysfunction associated with human islet amyloid
polypeptide overexpression. PloS One. 2014;9(7):e101797.
46

88. Rosas PC, Nagaraja GM, Kaur P, Panossian A, Wickman G, Garcia LR, et al. Hsp72
(HSPA1A) Prevents Human Islet Amyloid Polypeptide Aggregation and Toxicity: A New
Approach for Type 2 Diabetes Treatment. PloS One. 2016;11(3):e0149409.

47

Chapter 2
Diabetic Risk Factors Promote Islet Amyloid Polypeptide Misfolding by a
Common, Membrane-mediated Mechanism

Alan K. Okada
1
, Kazuki Teranishi
1
, J. Mario Isas
1
, Sahar Bedrood
2
, Robert H. Chow
3*
, Ralf
Langen
1*
1
Department of Biochemistry and Molecular Biology, Zilkha Neurogenetic Institute, University
of Southern California, Los Angeles, California, USA,
2
USC Eye Institute, Department of
Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles,
California, USA,
3
Department of Physiology and Biophysics, Keck School of Medicine, Zilkha
Neurogenetic Institute, University of Southern California, Los Angeles, California, USA
*
Corresponding Authors: Ralf Langen, Zilkha Neurogenetic Institute, Department of
Biochemistry and Molecular Biology, University of Southern California, Los Angeles, California
90033, USA. Tel.: 323-442-1323 Fax: 323-442-4404. Email: Langen@usc.edu R. H. Chow:
Zilkha Neurogenetic Institute, Department of Physiology and Biophysics, Keck School of
Medicine, University of Southern California. Los Angeles, CA 90089-2821, USA.
Email: rchow@usc.edu

48

ABSTRACT
The current diabetes epidemic is associated with a diverse set of risk factors including obesity and
exposure to plastics. Notably, significant elevations of negatively charged amphiphilic molecules
are observed in obesity (e.g. free fatty acids and phosphatidic acid) and plastics exposure
(monophthalate esters). It remains unclear whether these factors share pathogenic mechanisms and
whether links exist with islet amyloid polypeptide (IAPP) misfolding, a process central to β-cell
dysfunction and death. Using a combination of fluorescence, circular dichroism and electron
microscopy, we show that phosphatidic acid, oleic acid, and the phthalate metabolite MBzP
partition into neutral membranes and enhance IAPP misfolding. The elevation of negative charge
density caused by the presence of the risk factor molecules stabilizes a common membrane-bound
α-helical intermediate that, in turn, facilitates IAPP misfolding. This shared mechanism points to
a critical role for the membrane-bound intermediate in disease pathogenesis, making it a potential
target for therapeutic intervention.
Understanding the dysfunction and death of pancreatic β-cells in type 2 diabetes mellitus (T2DM)
is critical for developing disease-modifying therapeutics. Misfolding of islet amyloid polypeptide
(IAPP) is responsible in large part for the β-cell defect, but little is known about the interplay
between IAPP misfolding and the risk factors associated with T2DM. We find that IAPP
misfolding is enhanced by risk factors associated with obesity and phthalate exposure. All of these
factors act through a common pathway characterized by a meta-stable α-helical intermediate.
Targeting this intermediate may be of therapeutic relevance, especially when obesity or phthalate
exposure are factors.

49

2.1 INTRODUCTION
The misfolding and aggregation of the 37-amino acid peptide, islet amyloid polypeptide (IAPP),
is thought to be one of the main factors responsible for the β-cell defect that drives the pathogenesis
of type 2 diabetes mellitus (T2DM). More than 90% of patients with T2DM have IAPP aggregates,
and misfolded IAPP is toxic to cell lines. Transgenic animals expressing the aggregation-prone
human IAPP develop T2DM symptoms (1)(2)(3)(4). Epidemiological studies show that the
growing diabetes epidemic is correlated with an epidemic of obesity/metabolic syndrome, and to
exposure to plastics containing phthalates, a family of plastic additives known as plasticizers.
Whether metabolic or phthalate-based risk factors act independently of IAPP or exert their effects
by affecting IAPP misfolding is unknown.
Interestingly, both plasticizer exposure and metabolic syndrome lead to significant elevations of
negatively charged amphiphilic lipid or lipid-like molecules (Figure S1). For example, rodent
models of obesity have shown that levels of the negatively charged lipid, phosphatidic acid (PA),
are elevated in the insulin- and IAPP-producing pancreatic β-cells (5)(6). Moreover, insulin
resistance, obesity and dyslipidemia result in remarkably high levels of free fatty acids (FFAs) that
can reach up to millimolar concentrations (7)(8)(9)(10). On the other hand, exposure to certain
everyday plastics leads to elevated blood and urine levels of negatively charged phthalate
metabolites, known to be associated with a decrease in β-cell function and T2DM (11)(12)(13)(14).
These amphiphilic, negatively charged molecules are generated as metabolic by-products of
phthalate-based plasticizers, which are added to plastics in order to increase flexibility. Exposure
to plasticizers occurs under many different circumstances, as plasticizers leach out of many
common materials. For phthalate-based plasticizers, these include food packaging, furniture, toys,
50

medical devices, cosmetics, perfumes, lotions, paints, lacquers, varnishes, and pharmaceuticals
(11). Thus, a better understanding of their potential toxic effects is needed.
The central question addressed in the present study is whether the aforementioned, negatively
charged lipids or lipid-like molecules promote IAPP misfolding. If so, this would indicate that
metabolic and phthalate-based risk factors could contribute to T2DM pathogenesis by enhancing
IAPP toxicity. Prior work has shown that negatively charged membranes containing
phosphatidylserine or phosphatidylglycerol can promote IAPP misfolding (15)(16). This
membrane-mediated misfolding pathway involves an attractive interaction between the positively
charged IAPP and the negatively charged membrane surface, which results in a transient α-helical
structure of IAPP that catalyzes misfolding. Here we wanted to test whether PA, FFAs and
phthalate metabolites (monophthalate esters) might similarly be able to facilitate misfolding via a
membrane-mediated misfolding pathway. While the lipid PA is naturally found in membranes,
FFAs and monophthalate esters will insert their hydrophobic regions into phospholipid
membranes, leaving their negatively charged carboxyl groups at the membrane surface. By
embedding into the membrane, these molecules would enhance the negative charge density at the
membrane surface and thereby promote IAPP misfolding.
In order to address this question, we chose oleic acid (OA) as an example for a FFA, and we used
monobenzyl ester phthalate, MBzP, as an example of a monophthalate ester (Figure 2.S1). OA
was chosen because it is the most abundant FFA in pancreatic fat and in serum. Importantly, mice
fed on a high fat diet show significant increases in pancreatic free-OA and triacylglycerol–
associated OA. Furthermore, patients with higher pancreatic fat show a similar trend (17)(18).
MBzP was of particular interest because it is known to be associated with T2DM risk
(11)(12)(13)(14). Our data indicate that OA and MBzP strongly partition into membranes. Using
51

circular dichroism (CD), fluorescence spectroscopy and transmission electron microscopy (EM)
we also found that physiologically relevant amounts of PA, OA and MBzP potently promote IAPP
misfolding via a membrane-bound α-helical intermediate. Collectively, these data directly link
T2DM risk factors to IAPP misfolding via a common catalytic pathway.
Figure 2.S1

Figure 2.S1 – Cartoon schematic of the structures of A) POPA B) OA C) MBzP.

52

2.2 RESULTS
2.2.1 Phosphatidic Acid-LUVs accelerate IAPP misfolding
In order to test whether the negatively charged phospholipid PA can modulate IAPP misfolding
via a membrane-mediated misfolding mechanism, we used thioflavin T (ThT) fluorescence. IAPP
was incubated with PA-LUVs, large unilamellar vesicles (LUVs) composed of varying ratios of
phosphatidic acid (PA) and the neutral phospholipid phosphatidylcholine (PC). IAPP achieved
half-maximal fluorescence (t50) after approximately 9.5 h in the presence of PC-only vesicles (i.e.
100% PC LUV) (Figure 1b). All vesicles containing PA accelerated IAPP misfolding.
Remarkably, even 1 mol % PA, which corresponds to a substoichiometric concentration of PA
molecules (1 PA:2.5 IAPP), sped up misfolding by almost 4 hours (Figure 2.1). The strongest
acceleration of misfolding was observed for 10 and 25% PA where misfolding occurred nearly 20
times faster than in the control. This robust acceleration was mildly attenuated in the presence of
66 mol % PA. Overall, the behavior is reminiscent of what we observed for negatively charged PS
membranes, where the maximal enhancement of misfolding was observed at 25% PS and where
higher amounts of PS progressively reduced the kinetics of misfolding (15).
Interactions between the negatively charged PA and the positively charged IAPP are likely
electrostatic in nature (19). To test this notion, we examined the effect of increased ionic strength
on misfolding by repeating the experiments of Fig.1 in the presence of salt (100 mM NaCl). We
found that PA-LUVs retain their ability to accelerate IAPP misfolding under these conditions, but
that the optimum acceleration of misfolding occurs at a higher percentage of PA (Figures 2.S2 and
2.S3). A similar effect was observed when IAPP misfolding was triggered by PS (20). These data
suggest that, just as in the case of PS membranes, salt shields the electrostatic interactions between
the negative membrane and the positive peptide.
53

Figure 2.1

Figure 2.1 – Phosphatidic acid modulates the rate of IAPP fibrilization. IAPP misfolding was
measured as a function of time using ThT and demonstrates a dependence upon the mol % of PA.
A) Representative ThT curves from experiments with 12.5 μM human IAPP and 500μM l lipid in
10mM phosphate buffer pH 7.4 during incubation with LUVs composed of 66% PA (omitted in
A), 25% PA (□), 10% PA (Δ), 1% PA (X), and 0% PA (100% PC, ○). B) Comparison of average
t50 values with the mol % PA from experiments in A. Error bars represent one standard deviation
from a minimum of 3 experiments per condition. (p < 0.05 for 1%, and p < 0.01 for all other
conditions, compared with 100% PC-LUV control).

54

Figure 2.S2

Figure 2.S2 – Increased ionic strength right shifts the optimum PA content of vesicles for
acceleration of IAPP fibrilization. IAPP aggregation was measured as a function of time using
ThT in 10 mM phosphate buffer pH 7.4 with 100mM NaCl pH 7.4 during incubation with PA-
LUVs. A) Representative ThT curves from experiments with 12.5 μM human IAPP and 500 μM
composed of 66% POPA (omitted in A), 25% POPA (□), 10% POPA (Δ), 1% POPA(X), and
100% POPC (○). Normalization was performed as before (see Figure 1 or materials and methods).
B) Comparison of average t50 values with the mol % PA from experiments in A. Error bars
represent one standard deviation from a minimum of 3 experiments per condition. (p < 0.01, for
all conditions compared with 100% POPC LUV control).
55

Figure 2.S3

Figure 2.S3 – Increased ionic strength does not change IAPP fibril morphology, or association
with PA-LUVs by electron microscopy. 12.5 μM IAPP was allowed to misfold in the presence of
500 μM PA-containing LUVs. Samples were incubated for the amount of time required to achieve
ThT positivity before being applied to the micrograph grid. A – D) EM micrographs of IAPP in
10 mM phosphate buffer, 100 mM NaCl, pH 7.4 with 66 (a), 25 (b), 10 (c), and 1 mol% (d) PA-
LUVs. Similar to low salt conditions, these images show fibrils decorated with LUVs, in some
cases the circular shape of the LUVs are distorted where they contact fibrils. Lipid free sections of
fibrils measure diameters of ~7 nm. Where lateral assemblies of fibrils are evident, their apparent
diameters are wider than 7 nm. In addition, small spherical structures were often present and
ranged from 9 nm in diameter and greater (arrows). Scale bars = 200 nm.
2.2.2 IAPP transiently adopts an α-helical conformation with PA-LUVs before
transitioning to a β-sheet rich structure
We next set out to determine what structural changes occur as IAPP misfolds in the presence of
PA-LUVs. IAPP secondary structure was determined by CD at different time points throughout
the misfolding process. These experiments were limited to the low salt conditions as chloride
absorbs strongly in the lower UV range, which interferes with CD measurements. Immediately
after dissolution in buffer, IAPP displayed a negative peak near 202 nm (Figure 2.2 a-d, solid line),
56

characteristic of a mostly disordered structure. Upon addition of PA-LUVs, IAPP immediately
transitioned into an α-helical structure as indicated by the negative peaks at 208 and 222 nm
(Figure 2.2 a-c, dotted line). Within minutes to hours (depending upon conditions) the helical
structure then transitioned into a β-sheet rich conformation, indicated by the negative peak near
218 nm (Figure 2.2 a-c, dashed line). The initial degree of helicity increased with the PA content
of the vesicles (Figure 2.2 d). Using previously established methods (21)(22), we estimated the
percentage of helicity for IAPP in the presence of 66 mol % PA-LUVs to be approximately 45%.
Given that IAPP is 37 residues in length and assuming full binding, this estimate corresponds to
~17 amino acids participating in the membrane-bound helical structure. These findings are in line
with prior estimates from IAPP bound to membranes containing high percentages of PS (20).

57

Figure 2.2

Figure 2.2 – IAPP forms an α-helix upon interaction with phosphatidic acid containing LUVs
before transitioning into a β-sheet structure. A - C) Representative circular dichroism spectra are
shown. The secondary structure of 25 μM IAPP was first determined in 10mM phosphate buffer,
pH 7.4 (—). A minimum at 202 nm suggests a mostly disordered structure. LUVs composed of 66
(a), 25 (b), and 10 mol % (c) POPA were added and the structure was immediately measured (•••).
Minima at 208 nm and 222 nm indicate a transition into an α-helix upon interaction with LUVs.
After incubation at room temperature, the samples were measured a final time (— —). A minimum
near 218 nm is most consistent with a β-sheet structure. D) Initial spectra of IAPP upon addition
58

of 66 (— • —), 25 (— —), and 10 mol % (•••) PA-LUVs are overlaid with the disordered structure
in buffer (—), provided for reference.
2.2.3 Electron microscopy of IAPP incubated with PA-LUVs.
Collectively, the enhanced ThT fluorescence and β-sheet formation strongly suggested the
formation of IAPP fibrils. For additional validation, we visualized IAPP by negative-stain
transmission electron microscopy (EM) at late time points when the ThT readings had plateaued
(Figure 2.3). Indeed, fibrils could be observed in the EM. The morphology of fibrils formed in the
presence of PA-LUV was similar to those seen in the absence of lipids. Both cases yielded long,
non-branching fibrils that were often ribbon-like, sometimes twisted, and had diameters of ~7 nm.
In some areas, it appeared as if lateral assemblies of individual fibril strands thickened the apparent
diameter of fibrils. Interestingly, where lipids are present and contacting the surface of the fibrils,
these lipids are often distorted and also contribute to an apparent thickening of the fibril diameter
as in figure 2.3. In nearly all cases, fibrils can be found free of PA-LUVs. In addition, small
rounded structures were formed when IAPP and PA-LUVs were incubated. Typically, their sizes
ranged from ~ 6 to 20 nm in diameter. These structures could represent oligomeric IAPP, small
protein-lipid particles as was observed for a-synuclein (23)
,
(24) or both.
Taken together these data suggested that PA can accelerate misfolding in a manner similar to PS,
further supporting the notion that negative membrane charge, rather than the specific nature of the
lipids, plays a key role in facilitating membrane-mediated misfolding.
59

Figure 2.3

Figure 2.3. Electron microscopy reveals IAPP fibrils in some cases decorated with PA-LUVs. 12.5
μM IAPP was allowed to misfold in the presence of 500 μM PA-containing LUVs. Samples were
incubated for the amount of time required to achieve ThT positivity, as in figure 1, before being
applied to the micrograph grid. A – D) EM micrographs of IAPP in 10 mM phosphate buffer, pH
7.4 with 66 (a), 25 (b), 10 (c), and 1 mol % (d) PA-LUVs. These images show fibrils, many of
which are decorated with LUVs. In some cases the shape of the LUVs are distorted where they
contact fibrils. Lipid free sections of fibrils measure diameters of ~7 nm. Some fibrils assemble
laterally, increasing the apparent fibril diameter. Arrows indicate small spherical structures 6 nm
in diameter and greater (arrows). Scale bars = 200 nm.
2.2.4 Oleic Acid-LUVs accelerate IAPP misfolding kinetics
To test whether membrane-embedded amphiphilic molecules such as fatty acids can mimic
negatively charged lipids and enhance IAPP misfolding kinetics, we measured the ability of PC-
based LUVs containing the fatty acid, oleic acid (OA-LUV) (25)(26), to affect IAPP misfolding
(Figure 2.4). Using ThT fluorescence, we monitored IAPP misfolding in the presence of OA-LUVs
containing 1, 10, 25, and 66 mol % oleic acid. Compared to PA and PS (15), OA enhanced
misfolding even more. Particularly remarkable was the finding that even small substoichiometric
60

amounts of 1% OA (1 OA: 2.5 IAPP) reduced the t50 by nearly 7 hours. The fastest kinetics were
obtained for 66 mol % OA, where misfolding was enhanced by more than 100 fold. Given that OA
strongly partitions into PC membranes (27)(28), the enhanced misfolding was expected to be
caused by membrane-embedded, rather than soluble OA. Inasmuch as soluble OA has the capacity
to modulate IAPP misfolding (29), we further tested this notion using centrifugation of OA-LUVs.
The supernatant, which would contain any soluble OA, did not cause acceleration of IAPP
misfolding kinetics and did not change IAPP secondary structure according to CD (Figure 2.S4).
Taken together, these data show a robust and significant enhancement of IAPP misfolding kinetics
by all OA-LUV mol %’s tested that occurs at or on LUV membranes.
Figure 2.4

Figure 2.4 – Oleic Acid enhances IAPP misfolding kinetics. ThT was used to follow IAPP
aggregation kinetics. A) IAPP aggregation was measured as a function of time using ThT and
demonstrates dependence upon the mol % of oleic acid. A and B) Representative ThT curves from
experiments with 12.5 μM human IAPP and 500μM oleic acid in 10mM phosphate buffer pH 7.4
25% (□), 10% (Δ), 1% oleic acid (x), and 100% POPC (○). B) Comparison of t50 values with the
61

mol % of oleic acid in LUVs from experiments in A. Error bars represent one standard deviation
from a minimum of 3 experiments per condition. (p < 0.01 for all conditions compared to 100%
POPC LUV control).
Figure 2.S4

62

Figure 2.S4 – Aqueous phase OA and MBzP do not accelerate IAPP aggregation. Supernatants
taken from 25 mol % OA- and MBzP- LUV suspensions following ultracentrifugation added to
IAPP. A) CD measurements of 25 μM IAPP with (black) and without (grey) supernatant taken
from OA-LUV samples. B). CD measurements of IAPP with (black) and without (grey)
supernatant taken from MBzP-LUV samples. C) Average t50’s of 12.5 μM IAPP treated with
vehicle, OA-, or MBzP- LUV supernatants. Error bars represent +/- 1 standard deviation.
2.2.5 IAPP transiently adopts α-helical conformation with OA-LUVs before
transitioning into β-sheet structure
We next tested whether the acceleration of IAPP misfolding by OA-LUVs also proceeded via a
helical intermediate. We used CD to monitor changes in IAPP secondary structure during
incubation with 10, 25 and 66 mol % OA-LUVs (Figure 2.5). At initial time points immediately
after mixing, we observed an increase in helicity with increasing percentage of OA (Figure 2.5 c).
Over time, helical IAPP transitioned into a β-sheet-rich structure, as indicated by the single
negative peak between 218-220 nm (Figure 2.5 a-b).
Figure 2.5

63

Figure 2.5 – IAPP forms an α-helix upon interaction with oleic acid containing LUVs before
transitioning into a β-sheet structure. Representative circular dichroism spectra are shown. The
secondary structure of 25 μM IAPP was first determined in 10 mM phosphate buffer, pH 7.4 (solid
line). A minimum at 202 nm suggests a mostly disordered structure. LUVs composed of 66 (a)
and 25 mol % (b) oleic acid were added and the structure was measured again (dotted line). Minima
at 208 nm and 222 nm indicate a transition into an α-helix upon interaction with LUVs. After
incubation at room temperature, the samples were measured a final time (dashed line). A minimum
near ~218 nm is most consistent with a primarily β-sheet structure. C) Initial spectra of IAPP upon
addition of 66 (— • —), 25 (— —), and 10 mol % (•••) OA-LUVs are overlaid with the disordered
structure in buffer (—), provided for reference.
2.2.6 Electron microscopy of IAPP incubated with OA-LUVs
Negative-stain electron microscopy verified the formation of fibrils in the presence of OA-LUVs
(Figure 2.6). As in the aforementioned cases, fibrils were frequently in contact with vesicles.
Distortion of the vesicles can sometimes be observed and in those instances contributes to an
apparent widening of fibril diameters. Examples of fibrils not in contact with LUVs can be found
in all conditions. Again, small rounded structures could also be observed under these conditions.
Collectively, these data are very similar to those obtained for PA-LUVs and show that OA can
promote membrane-mediated misfolding in a manner akin to negatively charged phospholipids.
64

Figure 2.6

Figure 2.6 – Electron microscopy reveals OA-LUV mediated misfolding of IAPP yields IAPP
fibrils in some cases decorated with OA-LUVs. 12.5 μM IAPP was allowed to misfold in the
presence of 500 μM OA-LUVs. Samples were incubated for the amount of time required to achieve
ThT positivity, as in figure 1, before being applied to the micrograph grid. A – D) EM micrographs
of IAPP in 10 mM phosphate buffer, pH 7.4 with 66 (a), 25 (b), 10 (c), and 1 mol % (d) OA-LUVs
show fibrils decorated with LUVs, in some cases the shape of the LUVs are distorted where they
contact fibrils. Lipid free sections of fibrils measure diameters of ~7 nm. In many cases lateral
assemblies of fibrils are evident as are small spherical structures 5 nm in diameter (arrows). Scale
bars = 200 nm.
2.2.7 Monobenzyl ester phthalate-LUVs accelerate IAPP misfolding kinetics
Next we tested whether another negatively charged, amphiphilic molecule, the monophthalate
ester, monobenzyl ester phthalate (MBzP), could similarly mimic negatively charged lipids and
promote IAPP misfolding. Toward this end, we used ThT fluorescence to measure the misfolding
kinetics of IAPP in the presence of PC-based MBzP-LUVs (Figure 2.7). MBzP-LUVs significantly
enhanced the rate of IAPP misfolding. Again, we used centrifugation in order to verify that the
effects were caused by membrane-bound rather than soluble MBzP. As in the case of OA-LUVs,
we found that the vesicle fraction, but not the supernatant, was capable of enhancing aggregation
65

(Figure 2.S4). These data indicated that MBzP strongly partitioned into the bilayer, however, as
MBzP partitioning into PC bilayers had not been described in the literature, we determined the
partitioning of MBzP into PC membranes. First, we determined that the extinction coefficient for
MBzP at 237 nm is 6633.2 M
-1
cm
-1
(see materials and methods). We used this extinction
coefficient in order to quantify the amounts of MBzP free in solution and bound to the vesicles,
and arrived at a partitioning coefficient of 1.1 ± 0.7 x 10
-5
(for details see materials and methods).
This coefficient further demonstrated that the concentration of MBzP in solution relative to that in
the membrane is very low. Unfortunately, the strong UV absorption of MBzP-LUVs prohibited us
from using CD to determine the secondary structure of IAPP within the MBzP-LUV membranes.
Figure 2.7

Figure 2.7 – Monobenzyl ester phthalate-LUVs enhance IAPP misfolding kinetics. ThT was used
to follow IAPP aggregation kinetics. A) IAPP aggregation was measured as a function of time and
demonstrates dependence upon the mol % of MBzP. A) Representative ThT curves from
experiments with 12.5 μM human IAPP and 500μM MBzP-LUVs in 10mM phosphate buffer pH
7.4 demonstrate acceleration of fibrilization for 66 (omitted in A), 25 (□), 10 (Δ), 1 mol % MBzP
66

(x), and 100% POPC (○). B) Comparison of t50 values from experiments in A. Error bars represent
+/- 1 standard deviation from a minimum of 3 experiments per condition. (p < 0.01, for all
conditions compared with 100% POPC LUV control).
2.2.8 Electron microscopy of IAPP incubated with MBzP-LUVs.
We used negative-stain EM to visualize IAPP misfolding end-products following incubation with
MBzP-LUVs. Applying the same methodology as before (see figures 2.3 and 2.6), we observed
that all lipid conditions produced IAPP aggregates, the primary constituent of which were fibrils
(Figure 2.8). Furthermore, those fibrils were often physically associated with MBzP-LUVs. Where
this association occurs, distortion of the vesicle morphology can often be observed. Again, small
rounded structures were sometimes present. Collectively, these data demonstrate that MBzP
partitions strongly into membranes and enhances IAPP misfolding.
Figure 2.8

Figure 2.8 – Electron microscopy reveals IAPP fibrils in some cases decorated with MBzP-LUVs.
12.5 μM IAPP was allowed to misfold in the presence of 500 μM PA-containing LUVs. Samples
were incubated for the amount of time required to achieve ThT positivity, as in figure 1, before
being applied to the micrograph grid. A – D) EM micrographs of IAPP in 10 mM phosphate buffer,
pH 7.4 with 66 (a), 25 (b), 10 (c), and 1 mol % (d) MBzP-LUVs show fibrils decorated with LUVs.
Some LUVs are distorted where they contact fibrils. Lipid free sections of fibrils measure
67

diameters of ~7 nm. In many cases lateral assemblies of fibrils are evident as are small spherical
structures 10 nm in diameter and greater (arrows). Scale bars = 200 nm.

68

2.3 DISCUSSION
The current T2DM epidemic is fueled by a number of risk factors, making it important to
understand how these risk factors contribute to disease pathogenesis. Here we specifically
investigate three negatively charged, amphiphilic molecules associated with known metabolic and
phthalate-based risk factors. All of these molecules are currently considered to act independently
of IAPP misfolding. Metabolic abnormalities and obesity in particular are thought to drive T2DM
by promoting insulin resistance. In addition, metabolic dysfunction can lead to a heightened
inflammatory state that may also contribute to T2DM (30). Phthalate-based risk factors are
hypothesized to contribute to T2DM by altering fatty acid metabolism, lipid homeostasis and
adipogenesis via peroxisome proliferator-activated receptor (PPAR) activation (31)(32). Our data
indicate, however, that PA, OA and MBzP all have a pronounced effect on IAPP misfolding. All
three molecules partition into membranes where they enhance negative charge density and
promote membrane-mediated misfolding. These data strongly suggest that the risk factors can also
contribute to T2DM pathogenesis through an IAPP-dependent pathway.
The effects on IAPP misfolding required remarkably small amounts of negatively charged
molecules. In all cases, substoichiometric amounts were sufficient to strongly catalyze misfolding.
The most pronounced effects were observed for OA-containing membranes, where the addition of
1% OA increased misfolding more than 4 fold. Thus, even relatively small amounts of these
molecules can significantly affect IAPP misfolding. The emerging picture from protein misfolding
or amyloid diseases is that factors that promote misfolding enhance the likelihood of disease. In
contrast, inhibition of misfolding (33)(34), enhanced clearance or degradation of misfolded
proteins via heat shock proteins, proteasome or autophagy pathways tend to reduce pathology (35).
Typically, protein misfolding diseases are late in onset and it often takes years for imbalances in
69

protein homeostasis to lead to clinical symptoms. We, therefore, envisage a model in which the
three classes of molecules tested here adversely effect protein homeostasis by continually favoring
the misfolded state(s) of IAPP and ultimately cause disease.
How physiologically relevant are the concentrations used in our study? The concentrations for PA
have recently been reported in rodent obesity models. In these studies, the PA levels in membranes
were found to be in the range of ~ 3 to 5% mol %, while non-obese rodents had less than 2% PA
and cardiolipin combined (5). As shown in Figure 1, these percentages are in a range where
significant enhancements of IAPP misfolding can be observed. The precise concentrations of
MBzP and OA in cellular membranes are not known, but we can estimate that the concentrations
of these molecules can also likely reach the range tested in the present study. The strong
partitioning of MBzP into PC membranes suggests that significant amounts of monophthalate ester
must localize to cellular membranes. MBzP serum concentrations of ~0.14 μM (36) have been
reported. Based on the Kp determined in the present study, mol percentages of 1% MBzP could
easily be accomplished in membranes. As mentioned above, obesity can drive FA concentrations
into the millimolar range (8)(9)(10)(37)(38)(39) and the concentrations are potentially even higher
in the local environment of the fatty pancreas (7)(25)(40)(41)(42)(43)(44). Regardless of the
precise number it is clear that the FA concentrations are orders of magnitude higher than those of
MBzP and small percentages of FA, similar to those tested here, could readily accumulate in
cellular membranes.
Our data also indicate that the enhancement of membrane-mediated misfolding of IAPP is a rather
generic property of membranes that largely depends upon its negative charge density and to a
lesser extent on the exact chemical nature of the negatively charged lipid. In fact, this property
does not even require bona fide phospholipids and can be mimicked by negatively charged
70

amphiphilic molecules that can partition into the membrane. Moreover, in all cases tested
misfolding proceeded via an α-helical intermediate. This helical intermediate has been suggested
to anchor IAPP into the membrane where two-dimensional diffusion can rapidly promote
intermolecular interactions in a low dielectric environment that promotes secondary structure
formation, including misfolding to β-sheet structure (19)(45). Preventing the membrane-bound
helical structure from forming or stabilizing it to prevent further misfolding would be beneficial
in counteracting the effects of the risk factors driving high T2DM incidence rates. In fact, recent
work has shown that IAPP-helix stabilizing/targeting agents can prevent IAPP membraned-
mediated misfolding and cytotoxicity (46)(47)(48). Based on our data, we suspect that this
approach represents a promising avenue to mitigate the harm caused by risk factors like obesity
and phthalate exposure. Finally, membrane-mediated misfolding has been observed for amyloid-
β and α-synuclein (49). Thus, it may well be possible that analogous risk factors may also exist
that promote membrane-mediated misfolding in other amyloid diseases.

2.4 REFERENCES
1. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC. Beta-cell deficit and
increased beta-cell apoptosis in humans with type 2 diabetes. Diabetes. 2003 Jan;52(1):102–
10.
2. Ashcroft FM, Rorsman P. Diabetes Mellitus and the β Cell: The Last Ten Years. Cell. 2012
Mar 16;148(6):1160–71.
3. Hull RL, Westermark GT, Westermark P, Kahn SE. Islet amyloid: a critical entity in the
pathogenesis of type 2 diabetes. J Clin Endocrinol Metab. 2004 Aug;89(8):3629–43.
71

4. Haataja L, Gurlo T, Huang CJ, Butler PC. Islet amyloid in type 2 diabetes, and the toxic
oligomer hypothesis. Endocr Rev. 2008 May;29(3):303–16.
5. Rustenbeck I, Matthies A, Lenzen S. Lipid composition of glucose-stimulated pancreatic
islets and insulin-secreting tumor cells. Lipids. 1994 Oct;29(10):685–92.
6. Farese RV, DiMarco PE, Barnes DE, Sabir MA, Larson RE, Davis JS, et al. Rapid glucose-
dependent increases in phosphatidic acid and phosphoinositides in rat pancreatic islets.
Endocrinology. 1986 Apr;118(4):1498–503.
7. Taylor R. Pathogenesis of type 2 diabetes: tracing the reverse route from cure to cause.
Diabetologia. 2008 Oct;51(10):1781–9.
8. Reaven GM, Hollenbeck C, Jeng CY, Wu MS, Chen YD. Measurement of plasma glucose,
free fatty acid, lactate, and insulin for 24 h in patients with NIDDM. Diabetes. 1988
Aug;37(8):1020–4.
9. Hsu IR, Kim SP, Kabir M, Bergman RN. Metabolic syndrome, hyperinsulinemia, and
cancer. Am J Clin Nutr. 2007 Sep;86(3):s867-871.
10. Bergman RN, Kim SP, Hsu IR, Catalano KJ, Chiu JD, Kabir M, et al. Abdominal obesity:
role in the pathophysiology of metabolic disease and cardiovascular risk. Am J Med. 2007
Feb;120(2 Suppl 1):S3-8-32.
11. Centers for Disease Control and Prevention. Fourth National Report on Human Exposure to
Environmental Chemicals. Department of Health and Human Services; 2009. Report No.:
4.
72

12. James-Todd T, Stahlhut R, Meeker JD, Powell S-G, Hauser R, Huang T, et al. Urinary
phthalate metabolite concentrations and diabetes among women in the National Health and
Nutrition Examination Survey (NHANES) 2001-2008. Environ Health Perspect. 2012
Sep;120(9):1307–13.
13. Svensson K, Hernández-Ramírez RU, Burguete-García A, Cebrián ME, Calafat AM,
Needham LL, et al. Phthalate exposure associated with self-reported diabetes among
Mexican women. Environ Res. 2011 Aug;111(6):792–6.
14. Lind PM, Zethelius B, Lind L. Circulating levels of phthalate metabolites are associated
with prevalent diabetes in the elderly. Diabetes Care. 2012 Jul;35(7):1519–24.
15. Jayasinghe SA, Langen R. Lipid membranes modulate the structure of islet amyloid
polypeptide. Biochemistry (Mosc). 2005 Sep 13;44(36):12113–9.
16. Knight JD, Miranker AD. Phospholipid catalysis of diabetic amyloid assembly. J Mol Biol.
2004 Aug 27;341(5):1175–87.
17. Pinnick KE, Collins SC, Londos C, Gauguier D, Clark A, Fielding BA. Pancreatic ectopic
fat is characterized by adipocyte infiltration and altered lipid composition. Obes Silver
Spring Md. 2008 Mar;16(3):522–30.
18. Saifer A, Goldman L. The free fatty acids bound to human serum albumin. J Lipid Res.
1961 Jul 1;2(3):268–70.
19. Jayasinghe SA, Langen R. Membrane interaction of islet amyloid polypeptide. Biochim
Biophys Acta. 2007 Aug;1768(8):2002–9.
73

20. Jayasinghe SA, Langen R. Lipid membranes modulate the structure of islet amyloid
polypeptide. Biochemistry (Mosc). 2005 Sep 13;44(36):12113–9.
21. Luo P, Baldwin RL. Mechanism of helix induction by trifluoroethanol: a framework for
extrapolating the helix-forming properties of peptides from trifluoroethanol/water mixtures
back to water. Biochemistry (Mosc). 1997 Jul 8;36(27):8413–21.
22. Rohl CA, Baldwin RL. Comparison of NH exchange and circular dichroism as techniques
for measuring the parameters of the helix-coil transition in peptides. Biochemistry (Mosc).
1997 Jul 15;36(28):8435–42.
23. Varkey J, Mizuno N, Hegde BG, Cheng N, Steven AC, Langen R. α-Synuclein oligomers
with broken helical conformation form lipoprotein nanoparticles. J Biol Chem. 2013 Jun
14;288(24):17620–30.
24. Eichmann C, Campioni S, Kowal J, Maslennikov I, Gerez J, Liu X, et al. Preparation and
Characterization of Stable α-Synuclein Lipoprotein Particles. J Biol Chem. 2016 Feb 4;
25. Pinnick KE, Collins SC, Londos C, Gauguier D, Clark A, Fielding BA. Pancreatic ectopic
fat is characterized by adipocyte infiltration and altered lipid composition. Obes Silver
Spring Md. 2008 Mar;16(3):522–30.
26. Saifer A, Goldman L. The free fatty acids bound to human serum albumin. J Lipid Res.
1961 Jul 1;2(3):268–70.
27. Hamilton JA, Cistola DP. Transfer of oleic acid between albumin and phospholipid vesicles.
Proc Natl Acad Sci U S A. 1986 Jan;83(1):82–6.
74

28. Anel A, Richieri GV, Kleinfeld AM. Membrane partition of fatty acids and inhibition of T
cell function. Biochemistry (Mosc). 1993;32(2):530–6.
29. Ma Z, Westermark GT. Effects of free fatty acid on polymerization of islet amyloid
polypeptide (IAPP) in vitro and on amyloid fibril formation in cultivated isolated islets of
transgenic mice overexpressing human IAPP. Mol Med Camb Mass. 2002 Dec;8(12):863–
8.
30. Yaturu S. Obesity and type 2 diabetes. J Diabetes Mellit. 2011;1(4):79–95.
31. Feige JN, Gelman L, Rossi D, Zoete V, Métivier R, Tudor C, et al. The endocrine disruptor
monoethyl-hexyl-phthalate is a selective peroxisome proliferator-activated receptor gamma
modulator that promotes adipogenesis. J Biol Chem. 2007 Jun 29;282(26):19152–66.
32. Hurst CH, Waxman DJ. Activation of PPARalpha and PPARgamma by environmental
phthalate monoesters. Toxicol Sci Off J Soc Toxicol. 2003 Aug;74(2):297–308.
33. Hammarström P, Wiseman RL, Powers ET, Kelly JW. Prevention of Transthyretin Amyloid
Disease by Changing Protein Misfolding Energetics. Science. 2003 Jan 31;299(5607):713–
6.
34. Coelho T, Maia LF, Silva AM da, Cruz MW, Planté-Bordeneuve V, Lozeron P, et al.
Tafamidis for transthyretin familial amyloid polyneuropathy A randomized, controlled trial.
Neurology. 2012 Aug 21;79(8):785–92.
35. Hipp MS, Park S-H, Hartl FU. Proteostasis impairment in protein-misfolding and -
aggregation diseases. Trends Cell Biol. 2014 Sep;24(9):506–14.
75

36. Olsén L, Lampa E, Birkholz DA, Lind L, Lind PM. Circulating levels of bisphenol A (BPA)
and phthalates in an elderly population in Sweden, based on the Prospective Investigation
of the Vasculature in Uppsala Seniors (PIVUS). Ecotoxicol Environ Saf. 2012 Jan
1;75:242–8.
37. Hücking K, Hamilton-Wessler M, Ellmerer M, Bergman RN. Burst-like control of lipolysis
by the sympathetic nervous system in vivo. J Clin Invest. 2003 Jan;111(2):257–64.
38. Kim SP, Catalano KJ, Hsu IR, Chiu JD, Richey JM, Bergman RN. Nocturnal free fatty acids
are uniquely elevated in the longitudinal development of diet-induced insulin resistance and
hyperinsulinemia. Am J Physiol Endocrinol Metab. 2007 Jun;292(6):E1590-1598.
39. Shimabukuro M, Zhou YT, Levi M, Unger RH. Fatty acid-induced beta cell apoptosis: a
link between obesity and diabetes. Proc Natl Acad Sci U S A. 1998 Mar 3;95(5):2498–502.
40. Tushuizen ME, Bunck MC, Pouwels PJ, Bontemps S, van Waesberghe JHT, Schindhelm
RK, et al. Pancreatic fat content and beta-cell function in men with and without type 2
diabetes. Diabetes Care. 2007 Nov;30(11):2916–21.
41. Ou H-Y, Wang C-Y, Yang Y-C, Chen M-F, Chang C-J. The association between
nonalcoholic fatty pancreas disease and diabetes. PloS One. 2013;8(5):e62561.
42. Seppälä-Lindroos A, Vehkavaara S, Häkkinen A-M, Goto T, Westerbacka J, Sovijärvi A,
et al. Fat accumulation in the liver is associated with defects in insulin suppression of
glucose production and serum free fatty acids independent of obesity in normal men. J Clin
Endocrinol Metab. 2002 Jul;87(7):3023–8.
76

43. Lim EL, Hollingsworth KG, Aribisala BS, Chen MJ, Mathers JC, Taylor R. Reversal of
type 2 diabetes: normalisation of beta cell function in association with decreased pancreas
and liver triacylglycerol. Diabetologia. 2011 Oct;54(10):2506–14.
44. Pappan KL, Pan Z, Kwon G, Marshall CA, Coleman T, Goldberg IJ, et al. Pancreatic beta-
cell lipoprotein lipase independently regulates islet glucose metabolism and normal insulin
secretion. J Biol Chem. 2005 Mar 11;280(10):9023–9.
45. Apostolidou M, Jayasinghe SA, Langen R. Structure of alpha-helical membrane-bound
human islet amyloid polypeptide and its implications for membrane-mediated misfolding. J
Biol Chem. 2008 Jun 20;283(25):17205–10.
46. Pithadia AS, Bhunia A, Sribalan R, Padmini V, Fierke CA, Ramamoorthy A. Influence of
a curcumin derivative on hIAPP aggregation in the absence and presence of lipid
membranes. Chem Commun Camb Engl. 2016 Jan 5;52(5):942–5.
47. Hebda JA, Saraogi I, Magzoub M, Hamilton AD, Miranker AD. A peptidomimetic approach
to targeting pre-amyloidogenic states in type II diabetes. Chem Biol. 2009 Sep
25;16(9):943–50.
48. Daval M, Bedrood S, Gurlo T, Huang C-J, Costes S, Butler PC, et al. The effect of curcumin
on human islet amyloid polypeptide misfolding and toxicity. Amyloid Int J Exp Clin
Investig Off J Int Soc Amyloidosis. 2010 Sep;17(3–4):118–28.
49. Butterfield SM, Lashuel HA. Amyloidogenic protein-membrane interactions: mechanistic
insight from model systems. Angew Chem Int Ed Engl. 2010 Aug 2;49(33):5628–54.
77

50. Nielsen L, Khurana R, Coats A, Frokjaer S, Brange J, Vyas S, et al. Effect of environmental
factors on the kinetics of insulin fibril formation: elucidation of the molecular mechanism.
Biochemistry (Mosc). 2001 May 22;40(20):6036–46.
51. Huang C, Mason JT. Geometric packing constraints in egg phosphatidylcholine vesicles.
Proc Natl Acad Sci U S A. 1978 Jan;75(1):308–10.

2.5 ACKNOWLEDGEMENTS
Research reported in this publication was supported in whole or in part by the National Institutes
on Aging of the National Institutes of Health under award numbers AG027936 (RL) and
P50AG001542 (RHC). We would also like to acknowledge support of the NIGMS (GM85791 to
RHC) and NIDDK (DK60623 to RHC).
The content of this manuscript is solely the responsibility of the authors and does not necessarily
represent the official views of the National Institutes of Health.

2.6 AUTHOR CONTRIBUTION STATEMENT
R.L., R.H.C., and A.K.O. contributed to the conception and design of the project; A.K.O., K.T.
and S.B. contributed to data acquisition, analysis and interpretation; A.K.O., K.T., J.M.I., S.B.,
R.H.C. and R.L. participated in drafts, revisions and final approval of the manuscript being
submitted for publication.

78

2.7 COMPETING FINANCIAL INTERESTS STATEMENT
The authors have no competing financial interests to disclose.

2.8 ABBREVIATIONS
IAPP – islet amyloid polypeptide
MBzP – monobenzyl ester phthalate
T2DM – type 2 diabetes mellitus
POPA - 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphate
PA – phosphatidic acid
POPC - 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine
PC – phosphatidylcholine
FFA – free fatty acid
OA – oleic acid
LUV – large unilamellar vesicle
CD – circular dichroism
EM – electron microscopy
HFIP - Hexafluoro-2-propanol

79

2.9 MATERIALS AND METHODS
Materials. Hexafluoro-2-propanol (HFIP), and Oleic Acid (OA) were obtained from Sigma-
Aldrich (St. Louis, MO). Thioflavin T (ThT) was obtained from Sigma-Aldrich (Milwaukee, WI).
1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine (POPC) and 1-palmitoyl-2-oleoyl-sn-
glycero-3-phosphate (POPA) were obtained from Avanti Polar Lipids (Alabaster, AL) as stocks
concentrated in chloroform. Monobenzyl ester phthalate (MBzP) was obtained from Tokyo
Chemical Industry - America (Portland, OR). Synthetic wild-type human IAPP was obtained from
Bachem (Torrance, CA).
Preparation of IAPP. IAPP was received from Bachem in a lyophilized powder, dissolved in HFIP,
aliquoted into individual tubes, frozen in N2 (l) and lyophilized. Protein concentrations were
determined by UV absorbance at 280 nm in 8M guanidinium chloride using an extinction
coefficient of 1405 M
-1
cm
-1
and verified by CD spectroscopy upon resolublization. Lyophilized
IAPP stocks were stored, desiccated at -80°C.
Preparation of Large Unilamellar Vesicles (LUV). Lipid vesicles were made by combining the
indicated molar ratios of PC with either PA, OA or MBzP dissolved in chloroform or ethanol,
evaporated to a dry film with N2 (g) and vacuum desiccated overnight. All PC was of the POPC
type while all PA was of the POPA type. The dried lipids were reconstituted in the appropriate
buffer and freeze-thawed at least 6 times before extrusion through a mini-extruder containing a
polycarbonate filter to a diameter of 100nm. For OA- and MBzP- LUVs, a centrifugation step
(55,000*g for 30 minutes) was added prior to extrusion to eliminate potential micellar
contaminants. The supernatant was removed following centrifugation and replaced with
appropriate buffer.
80

Thioflavin T Fluorescence Assay. A 5 mM stock concentration of Thioflavin T (ThT) was stored
at -20°C in water and used at a 25μM final concentration as before to monitor IAPP misfolding
kinetics (20). IAPP aliquots were prepared as above. Individual samples of IAPP were solubilized
in appropriate buffer with ThT to an appropriate concentration from a dry powder. Prior to the
addition of lipid, the solution was loaded into a 2mm path length quartz cuvette. CD and ThT
readings were taken to verify the concentration and quality of the IAPP. At time = 0, LUVs were
added, the solution mixed, and place into a JASCO FP-6500 fluorometer. Fluorescence was
monitored under the following settings and conditions: excitation wavelength = 450 nm, emission
wavelength = 482 nm, excitation slit width = 1 nm, emission slit width = 10 nm, temperature =
room temperature, and pH = 7.4.
t50 values were determined as before (20). Briefly, ThT fluorescence emission intensity was
recorded as a function of time. The data was fitted to a sigmoidal curve given by equation (1)
(1) 𝐼 = ( 𝐼 𝑖 + 𝑚 𝑖 𝑡 )+
( 𝐼 𝑓 +𝑚 𝑓 𝑡 )
(1+𝑒 −(
𝑡 −𝑡 50
𝜏 )
)
,
where I is the ThT fluorescence intensity, t is time, and t50 is time at half-maximal fluorescence
intensity. All other parameters are as described previously (50). t50 values were obtained from the
fit to our data and were used to compare misfolding kinetics of the various experimental conditions.
Normalization was performed by dividing the fluorescence intensity by the maximum intensity
read following the end of each experiment.
Circular Dichroism Spectroscopy. IAPP was prepared as above. Individual aliquots were
solubilized in appropriate buffer and transferred to a 1 or 2 mm quartz cuvette. CD spectra were
measured between 195 and 260 nm in a JASCO 815 spectropolarimeter. Measurements were taken
81

every 1 nm at a scan rate of 50 nm/min, with an averaging time of 1 s and background subtracted
using appropriate backgrounds.
The fraction of helicity was determined using the previously determined relationships (21)(22)
described in equations (2) and (3)
(2) 𝑓 𝛼𝐻
=
( 𝜃 𝑜𝑏𝑠 −𝜃 𝑅𝐶
)
( 𝜃 𝐻 −𝜃 𝑅𝐶
)
,
where fαH is the fraction of helicity, θobs is the observed ellipticity, θRC is the ellipticity value for a
fully random coiled peptide equal in length to IAPP, and θH is the ellipticity value for a fully helical
peptide equal in length to IAPP. Ellipticity values were taken at 222 nm and converted into mean
residual ellipticity using equation (3)
(3) 𝜃 =
𝜃 𝑟𝑎𝑤 𝑛𝐶𝑙
,
where θ is the mean residual ellipticity (deg*cm
2
dmol
-1
), θraw is the ellipticity in millidegrees, n
is the number of amino acids, C is the molar concentration, and l is the path length in millimeters.
The relationship described by equation (2) is only valid for proteins passing through a helix-coil
transition in conditions of constant temperature. Our experiments satisfy these conditions. In
addition, the accuracy of the estimate generated by this equation depends in part on the values used
to represent fully helical or random coil structures. As in our previous study, we used a value of
θH = -34.7 x 10
3
deg cm
2
dmol
-1
and θRC =
-
1.5 x 10
3
deg cm
2
dmol
-1
to describe fully helical and
random coil structures, respectively (20). Finally, despite using excess lipid, we cannot rule out
the possibility that some small fraction of IAPP remains unbound and must treat estimates of
helicity with caution.
82

Determination of MBzP extinction coefficient. MBzP was solubilized at concentrations between
35.7 ng/mL – 84.5 μg/mL in 10mM phosphate buffer, pH 7.4. UV-Vis spectra were obtained in a
1 cm quartz cuvette on a Beckman DU-640 spectrometer from 220 nm – 340 nm. Appropriate
backgrounds were subtracted. Absorption at 237 nm was recorded and a best-fit linear regression
was plotted according to the Beer-Lambert law: A = 6633.2C - 0.0007, R
2
= 0.99. As the y-
intercept was negligible with respect to the slope, we defined the slope of the above equation as
the molar extinction coefficient. ε237 = 6633.2 M
-1
cm
-1
.
Determination of MBzP partitioning coefficient in PC bilayers. The partitioning coefficient of
MBzP into PC-based bilayers was determined using previously developed methods (28) according
to equation (4)
(4) 𝐾 𝑝 =
[𝑀𝐵𝑧𝑃 ]
𝑚 [𝑀𝐵𝑧𝑃 ]
𝑎 ,
where Kp is the partitioning coefficient of MBzP into PC-bilayers, [MBzP]m is the concentration
of MBzP in the membrane phase and [MBzP]a is the concentration of MBzP in solution. Kp,
expressed in measurable quantities is represented in equation (5)
(5) 𝐾 𝑝 = (
[𝑀𝐵𝑧𝑃 ]
𝑡𝑜𝑡 −[𝑀𝐵𝑧𝑃 ]
𝑎 [𝑀𝐵𝑧𝑃 ]
𝑎 )
𝑉 𝑎 𝑉 𝑚 ,
where [MBzP]tot is the added MBzP concentration, Va is the aqueous volume, and Vm is the
phospholipid volume. Vm was approximated using a bilayer depth of 37 Å and an area per
phospholipid of 69 Å
2
(51). The relationship elaborated by equation (4) was also used to estimate
the mol % MBzP based on literature values of MBzP serum levels.
Vesicles containing MBzP in these experiments were made as above except no extrusion step was
used in order to ensure no loss of MBzP in the determination of soluble MBzP values. Vesicles
83

were pelleted via centrifugation at 55,000 x g for 30 minutes and [MBzP] a was determined by
measuring the supernatant for absorbance at 237 nm. These data were background subtracted using
the appropriate buffer. Reported Kp value is the average of at least three experiments performed
using vesicles composed of 1, 10, 25, and 66 mol % MBzP.
Electron Microscopy Studies. 10 μL of sample were applied to carbon-coated Formvar films
mounted on copper grids and the excess liquid was blotted away prior to application of 1% uranyl
acetate negative sate. Specimens were imaged using a Jeol-1400 transmission electron microscope
operated at 100kV. Samples were only applied to the EM grids after enough time had elapsed for
IAPP to be fully fibrilized based on ThT studies.

84

Chapter 3
Cardiolipin Accelerates the Misfolding of Islet Amyloid Polypeptide
Alan K. Okada
1
, Kazuki Teranishi
1
Ralf Langen
1*
1
Department of Biochemistry and Molecular Biology, Zilkha Neurogenetic Institute, University
of Southern California, Los Angeles, California, USA
*
Corresponding Author: Ralf Langen, Zilkha Neurogenetic Institute, Department of
Biochemistry and Molecular Biology, University of Southern California, Los Angeles, California
90033, USA. Tel.: 323-442-1323 Fax: 323-442-4404. Email: Langen@usc.edu

85

ABSTRACT
Type 2 diabetes mellitus is characterized by a progressive decline in β-cell number and
responsiveness. Both mitochondrial dysfunction and the misfolding of the 37-amino acid peptide,
islet amyloid polypeptide, are emerging as critical pathogenic elements underlying this decline.
Recent work has demonstrated that IAPP in both rodents and humans accesses the mitochondrial
organelle, but insofar as membrane interactions with IAPP are thought to be a key step that triggers
IAPP misfolding, little is currently known about the nature of interactions between mitochondrial
lipids and IAPP. Here we have studied the effects of the mitochondrial lipid, cardiolipin, on IAPP
misfolding as well as the effects IAPP has on cardiolipin-containing membranes. Based on data
from fluorescence spectroscopy, transmission electron microscopy and circular dichroism we find
that not only do membranes containing cardiolipin accelerate IAPP misfolding via a membrane-
mediated mechanism characterized by the formation of an α-helical intermediate, but IAPP can
induce the remodeling of cardiolipin-containing membranes. These results suggest that IAPP and
cardiolipin exert influence upon each other at a structural level. It is likely that the enhancement
of IAPP misfolding by cardiolipin represents a pathologically relevant and mutually destructive
event. On the other hand, the importance of IAPP’s capacity to remodel cardiolipin-containing
membranes has yet to play out as a normal function of IAPP in mitochondria, or an early event in
mitochondrial dysfunction.

86

3.1 INTRODUCTION
Type 2 diabetes mellitus (T2DM) is characterized by the progressive failure of β-cells within the
pancreatic islets (1). Their inability to respond to metabolic demands associated with insulin
resistance ultimately leads to decompensation and clinical manifestations of T2DM. This is in part
due to changes in β-cell responsiveness to metabolic signals and in part due to loss of β-cells (1).
The progressive loss of β-cell mass with concomitant replacement by amyloid deposits composed
primarily of islet amyloid polypeptide (IAPP) (2) is associated with apoptotic cell death (3), a
process centralized in large part at the mitochondria (4). This process is accompanied by structural
and functional abnormalities in islet mitochondria in diabetic patients (5) and as such
mitochondrial dysfunction is now a recognized and well-documented early event in the death of
β-cells and T2DM (4).
The apoptotic process in T2DM can be recapitulated by β-cell expression of, or treatment with,
human IAPP (2). IAPP expression and treatment lead to the production of reactive oxygen species
(ROS) as well as defects in mitochondrial function (6–9). It is believed that the process of IAPP
misfolding generates species toxic to the β-cell that are ultimately responsible for the cytotoxicity
of the peptide (2,10). The emerging understanding is one wherein IAPP misfolding is driven by
inappropriate and deleterious membrane disruptive interactions with anionic phospholipids (11–
15). While mitochondria share some phospholipids in common with the cell (16), distinct among
the phospholipid family is the mitochondrial-specific phospholipid, cardiolipin. Cardiolipin
consists of two phosphatidic acid (see Figure 2.S1 for phosphatidic acid structure) moieties linked
covalently via a glycerol linker (16). Mitochondria from both rodent and human islets are
infiltrated by IAPP (17). Despite this, little is known regarding the interplay between IAPP and
mitochondrial lipids.
87

In order to better understand the consequences of IAPP and mitochondrial lipid interactions, we
have investigated the interaction between IAPP and the uniquely mitochondrial lipid, cardiolipin.
Using thioflavin T (ThT) fluorescence and circular dichroism, we have found that cardiolipin can
dramatically enhance the misfolding of IAPP via a membrane-mediated mechanism. Transmission
electron microscopy further demonstrates that cardiolipin containing membranes are often found
deformed around IAPP fibrils, and at early time points can be found to be tubulated. In this way
we show that IAPP and cardiolipin mutually affect structural rearrangements in each other that
may have significant consequences to T2DM disease pathogenesis.

88

3.2 RESULTS
3.2.1 Cardiolipin-LUVs accelerate IAPP misfolding
In order to explore whether cardiolipin could modulate the kinetics of IAPP misfolding, we
monitored the misfolding of IAPP as a function of time using thioflavin T (ThT) fluorescence in
the presence of cardiolipin. Large unilamellar vesicles (LUVs) composed of the anionic
phospholipid cardiolipin (CL) and the neutral phospholipid phosphatidylcholine (PC) were
prepared as before (18). Time to half maximal fluorescence (t50) of IAPP in the presence of PC-
only LUVs (i.e. 100% PC-LUV) was ~9.5 hours. In the presence of 25, 10 and 1 mol % cardiolipin-
containing vesicles (CL-LUVs) IAPP misfolding was dramatically enhanced. While all cardiolipin
containing vesicles tested accelerated IAPP misfolding, an apparent inversion of dose dependency
was observed (Figure 3.1). In the case of phosphatidylserine and phosphatidic acid charged LUVs,
increasing anionic phospholipid content leads to an optimal mol % for inducing fibrilization of
IAPP and higher charge density slows the kinetics of misfolding (Figure 2.1 and (18)). In this case
it is possible that the optimal charge density is at or below the 1 mol % level, explaining the
apparent contradiction imposed by the inversion. In addition, a biphasic growth curve is present
when IAPP is treated with 1 mol % CL-LUVs. This suggests that there are potentially two distinct
modes of aggregation in this condition, a fast and slow reaction. Interestingly, transmission
electron micrographs taken during the early phase of misfolding display both fibril elongation
extending outward from the borders of LUVs and tubulation of CL-LUVs (Figure 3.2 A, B).
Electron densities of 4 – 5 nm consistent with phospholipid bilayer thicknesses were observed in
figure 3.2B-D, differentiating these lipid tubule structures from fibrils. These images demonstrate
that IAPP has the capacity to remodel cardiolipin-containing membranes by inducing curvature.

89

Figure 3.1

Figure 3.1 – IAPP incubated with CL-LUVs reveals dramatic enhancement of misfolding kinetics.
Misfolding of IAPP as a function of time monitored with ThT demonstrates an inverse dependence
upon the mol % of cardiolipin. A) Representative traces from experiments with 12.5 μM IAPP and
500 μM lipid in 10 mM phosphate buffer pH 7.4 during incubation with LUVs composed of 25%
CL (□), 10% CL (Δ), 1% CL (X), and 0% CL (100% PC, ○). B) Averaged t50 values from
experiments in A as a function of the mol % CL. For 1% CL-LUV treated IAPP, the first (fast)
phase of the biphasic curve is plotted in B. There is no significant difference between 1% and 10%
when comparing the second phase of the biphasic curve with 10% CL-LUV treated samples. Error
90

bars represent one standard deviation from a minimum of 3 experiments per condition. (p < 0.01
compared with 100% PC-LUV control).
Figure 3.2

Figure 3.2 – IAPP extends fibrils from the surface of vesicles and remodels cardiolipin membranes
into tubules. A) 12.5 µM IAPP incubated for 30 minutes with 500 µM, 1 mol % CL-LUVs and
applied to micrograph grids and visualized by Transmission Electron Microscopy. B-D) Tubules
formed from membranes composed of B) 1 mol % CL-LUV, C) 25 mol % CL-LUV, and D) 20
mol % CL-MLV upon 30-minute incubation with IAPP. Arrowheads indicate amyloid fibrils.
Arrows indicate tubules. Scale bars = 200 nm.

91

3.2.2 Physiological salt concentrations rectify dose-dependent acceleration of IAPP
misfolding
In order to determine whether electrostatic interactions (19) between the anionic cardiolipin and
cationic IAPP are, at least in part, responsible for the interplay between the two, we studied the
consequences of increased ionic strength on the misfolding of IAPP. IAPP was incubated with 25,
10 and 1 mol % CL-LUVs in 10mM phosphate buffer, pH 7.4 with 100 mM NaCl and monitored
for misfolding by ThT fluorescence. We found that incubation with CL-LUVs in these conditions
again enhanced the misfolding kinetics of IAPP (Figure 3.3). Interestingly, under these conditions
the dose-dependency of IAPP misfolding rate enhancement was rectified such that higher charge
densities accelerated IAPP misfolding more strongly than lower charge densities. This behavior is
in line with observations of IAPP incubated with PS and PA ((18), Figure 2.S1) in similar
conditions. Based upon the shielding effects of salt, the data indicate that electrostatic interactions
play an important role in the association between cardiolipin and IAPP.
92

Figure 3.3

Figure 3.3 – Increased ionic strength rectifies the apparent inversion of dose-dependence for
acceleration of IAPP fibrilization. IAPP misfolding as a function of time monitored by ThT
fluorescence in 10 mM phosphate buffer pH 7.4 with 100mM NaCl during incubation with CL-
LUVs. A) Representative ThT measurements of12.5 μM IAPP and 500 μM LUVs composed of
25% CL (□), 10% CL (Δ), 1% CL (X), and 0% CL (100% PC, ○). B) Averaged t50’s from (A) as
a function of the mol % CL. Error bars show one standard deviation from a minimum of 3
experiments per condition. (p < 0.01 compared with 100% PC-LUV control).

93

3.2.3 IAPP undergoes structural transitions through α-helical conformations upon
interaction with Cardiolipin-LUVs
In order to test whether interactions with cardiolipin induce structural changes in IAPP, circular
dichroism (CD) was used to monitor secondary structure at time points during the misfolding
process. Due to the absorption of lower UV range light by chloride, low salt conditions were used
in these experiments to prevent interference with measurements. Measurements were taken
immediately upon solubilization of IAPP in buffer in order to verify IAPP concentration and
quality. As expected, IAPP in solution had a negative peak near 202 nm, consistent with a peptide
that is mostly disordered in solution (Figure 3.4 - solid line). Measurements were retaken following
addition of 500 µM CL-LUVs. CL-LUVs induced helicity in IAPP as demonstrated by the
increased negative signal at 208 and 222 nm (Figure 3.4 A, B - dotted line). Samples were
incubated at room temperature and after minutes to hours, shifted structural conformations to a
predominantly β-sheet ordering based on the development of a peak near 218 nm (Figure 3.4 A, B
- dashed line). The degree of helicity induced by CL-LUVs was dependent upon the mol %
cardiolipin. This was evident both at 1:20 protein to lipid ratio as well as in conditions where lipid
was added in excess to saturate binding to CL-LUVs (Figure 3.4 C). In these conditions the
percentage of IAPP helicity can be estimated using previously defined methods (20)(21). Based
on the amplitude of mean residual ellipticity at 222 nm, we calculated the percentage of helicity
for IAPP bound to 25% CL-LUVs to be approximately 29%. For a 37 amino acid peptide, this
indicates that 11 amino acids form the helical segment of IAPP in the cardiolipin-membrane bound
state.
.

94

Figure 3.4

Figure 3.4 – IAPP forms an α-helix upon interaction with cardiolipin containing LUVs prior to a
β-sheet transition. A - B) Representative circular dichroism spectra. Secondary structure of 25 µM
IAPP was determined in 10mM phosphate buffer, pH 7.4 (—). Minima at 202 nm suggests a
mostly disordered structure upon solubilization. 500 µM LUVs composed of 25 (a) and 10 (b) mol
% cardiolipin were added and measured immediately (•••). A transition into an α-helix upon
interaction with LUVs is indicated by minima at 208 nm and 222 nm. The samples were measured
a final time after incubation at room temperature (— —). A single minimum at or near 218 nm
indicates β-sheet structuring. C) Spectra of IAPP upon addition of saturating amounts of 25 (•••)
and 10 (— —) mol % CL-LUVs are shown. IAPP in solution (—) is given as reference.
3.2.4 Transmission Electron Microscopy of IAPP incubated with Cardiolipin-LUVs
In order to confirm the results of the fluorescence and CD studies, we visualized the interactions
between IAPP and CL-LUVs using negative stain transmission electron microscopy. Samples
taken in the plateau phase of the ThT fluorescence experiments were applied to TEM grids. We
observed fibrils in all CL-LUV treated conditions (Figure 3.5). Morphologically, fibrils formed in
the presence of cardiolipin were non-branching and ~7 nm in diameter where association with
95

LUVs or other fibrils do not obscure measurement. In fact, many of the fibers are heavily decorated
with vesicles in all conditions and lateral assemblies of fibers can often be seen as is demonstrated
in figure 3.4A. Collectively, these data indicate that IAPP interacts with CL-LUVs and this
interaction ultimately drives fibrilization.
Figure 3.5

Figure 3.5 – Transmission electron microscopy shows IAPP decorated with CL-LUVs. IAPP
incubated in the presence of 500 μM CL-LUVs for enough time to achieve ThT positivity (see
figure 1), then applied to the micrograph grid. A – C) TEM micrographs of IAPP in 10 mM
phosphate buffer, pH 7.4 with 25 (a), 10 (b), and 1 mol % (c) CL-LUVs. Many of the fibrils in
these conditions are observed decorated with CL-LUVs. LUVs in contact with fibrils are often
distorted or small. Fibrils are ~7 nm in diameters. Scale bars = 200 nm.

96

3.3 DISCUSSION
Mitochondrial dysfunction within β-cells is a well-documented and early component of T2DM,
yet the causes of mitochondrial dysfunction in T2DM are poorly understood. Here we have
investigated a potential link between IAPP misfolding and mitochondria by examining the
consequences of IAPP interaction with the uniquely mitochondrial lipid cardiolipin. Collectively,
our data show that when IAPP interacts with cardiolipin, the misfolding rate of IAPP is enhanced
dramatically through a membrane-mediated misfolding mechanism. By monitoring IAPP
misfolding with thioflavin T, we found that changes in IAPP misfolding kinetics are dependent
upon the dose of cardiolipin, as well as the ionic strength of the environment. Changes in ionic
strength had strong effects upon the misfolding kinetics of IAPP in the presence of cardiolipin,
suggesting that electrostatic forces between IAPP and cardiolipin are fundamentally important to
the interaction. IAPP is positively charged at the N-terminus, K1, R11, and depending on pH, H18.
Cardiolipin carries two negatively charged phosphate headgroups tethered together. As such,
increased ionic strength likely shields interactions between the charged groups on IAPP and
cardiolipin, altering the misfolding kinetics of IAPP.
Data from circular dichroism spectroscopy demonstrates that cardiolipin rapidly induces helicity
in IAPP which then undergoes a slower transition to β-sheet. When incubated with 25 mol % CL-
LUVs, the highest concentration tested, the helicity correlated to
~
11 amino acids, an estimate in
line with models of IAPP helicity determined from EPR spectroscopy (22), as well as phosphatidic
acid-LUVs at 25 mol %, for which we estimate ~32% helicity, or 12 amino acids participating in
the helix. Thus, cardiolipin drives the membrane-mediated misfolding of IAPP. This result further
highlights that the capacity to induce helicity in IAPP and drive its subsequent misfolding is a
general property of negatively charged membranes.
97

Electron microscopy of IAPP samples treated with CL-LUVs taken after the misfolding process
has reached its plateau phase (based on ThT fluorescence intensity changes), reveal that IAPP
fibrils are often decorated with CL-LUVs that in many cases are strongly distorted. In fact, EM
grids taken during the early phase of the misfolding process, show both tubulation of membranes
as well as what appears to be nascent fibril growth from the surfaces of vesicles. The latter effect
is consistent with primary nucleation events occurring on the membrane itself and in line with
previous findings with phosphatidylserine containing vesicles (14). Electron microscopy studies
of rodent and human β-cells reveals that IAPP does in fact access the mitochondrial compartment
(17). It is not clear whether the access itself is physiologically normal or a precursor to pathology.
In much the same way, it cannot be ruled out that the tubulating effects we observe in the early
stages of the IAPP-cardiolipin reaction represent physiologically relevant event. In fact, the rodent
analog of IAPP does not misfold or induce β-cell toxicity (23) and can be found within
mitochondria (17), raising the possibility that it performs some function there. The remodeling of
membranes by protein-membrane interaction, when regulated, can be a normal and indeed vital
cellular process underlying a variety of important functions (24). On the other hand, unregulated,
the remodeling of cellular membranes can disrupt membrane integrity (25). The ability of IAPP to
tubulate cardiolipin membranes is consistent with prior findings with IAPP and phosphatidylserine
membranes (17) wherein remodeling of membranes has been described. By this mechanism IAPP
helices insert shallowly into one face of a bilayer and spread the phospholipid headgroups,
inducing positive curvature. Cardiolipin, however, is considered a non-bilayer forming lipid due
to a mismatch between the relatively small size of the headgroup and the large hydrophobic tails,
which gives it a conical shape. Consequently, cardiolipin will tend to induce negative curvature in
phospholipid bilayers, especially when distributed asymmetrically as is the case in mitochondrial
98

membranes (16,26,27). A helical wedge inserted shallowly between cardiolipin headgroups can
alleviate strain induced by the intrinsic negative curvature generated by cardiolipin in an
asymmetric distribution. Conversely, helical wedge insertion on the outer face of a CL-LUV where
cardiolipin distribution is symmetrical between leaflets is expected to generate positive curvature,
consistent with the tubulation we observe. Ultimately, these notions and the known presence of
IAPP in the mitochondria, raise the question, could IAPP-induced mitochondrial membrane
remodeling be important for mitochondrial form and function, perhaps by wedging into cardiolipin
rich membranes to alleviate curvature strain?
An alternate possibility is that remodeling of cardiolipin membranes by IAPP are potentially
pathological. In fact, the current understanding of the mechanisms of IAPP toxicity centers around
the disruption of plasma and organelle membranes via fragmentation of the membrane or via pore
formation (10). Should disruption of membranes occur in the mitochondria there could be
significant consequences. Cardiolipin is found predominately within the inner mitochondrial
membrane reaching concentrations of greater than 24% (wt/wt). Within the inner mitochondrial
membrane, the majority of the cardiolipin is oriented facing the matrix side, but a minor fraction
can be found facing the inner mitochondrial space. In the outer mitochondrial membrane,
cardiolipin is less abundant, but faces primarily into the cytoplasm (16,27,28). This suggests that
IAPP-mediated disruption of cardiolipin membranes within mitochondria could occur on the outer
leaflet of the outer mitochondrial membrane, or on either leaflet of the inner mitochondrial
membrane. In either case, the electrochemical proton gradient stored in the inner membrane space
could be drained. In addition, interactions between IAPP and cardiolipin could lead to cytochrome
C release from cardiolipin, leading to apoptotic activation.
99

Future studies will be required to better understand the consequences of IAPP-cardiolipin
interactions and to determine whether they represent a physiological or pathological element in the
β-cell.
3.4 REFERENCES
1. Alejandro EU, Gregg B, Blandino-Rosano M, Cras-Méneur C, Bernal-Mizrachi E. Natural
history of β-cell adaptation and failure in type 2 diabetes. Mol Aspects Med. 2015
Apr;42:19–41.
2. Haataja L, Gurlo T, Huang CJ, Butler PC. Islet amyloid in type 2 diabetes, and the toxic
oligomer hypothesis. Endocr Rev. 2008 May;29(3):303–16.
3. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC. Beta-cell deficit and
increased beta-cell apoptosis in humans with type 2 diabetes. Diabetes. 2003 Jan;52(1):102–
10.
4. Ma ZA, Zhao Z, Turk J. Mitochondrial Dysfunction and β-Cell Failure in Type 2 Diabetes
Mellitus. Exp Diabetes Res [Internet]. 2012 [cited 2016 May 4];2012. Available from:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3216264/
5. Anello M, Lupi R, Spampinato D, Piro S, Masini M, Boggi U, et al. Functional and
morphological alterations of mitochondria in pancreatic beta cells from type 2 diabetic
patients. Diabetologia. 2005 Feb;48(2):282–9.
6. Huang C, Lin C, Haataja L, Gurlo T, Butler AE, Rizza RA, et al. High expression rates of
human islet amyloid polypeptide induce endoplasmic reticulum stress mediated beta-cell
100

apoptosis, a characteristic of humans with type 2 but not type 1 diabetes. Diabetes. 2007
Aug;56(8):2016–27.
7. Konarkowska B, Aitken JF, Kistler J, Zhang S, Cooper GJS. Thiol reducing compounds
prevent human amylin-evoked cytotoxicity. FEBS J. 2005 Oct;272(19):4949–59.
8. Li X-L, Xu G, Chen T, Wong Y-S, Zhao H-L, Fan R-R, et al. Phycocyanin protects INS-1E
pancreatic beta cells against human islet amyloid polypeptide-induced apoptosis through
attenuating oxidative stress and modulating JNK and p38 mitogen-activated protein kinase
pathways. Int J Biochem Cell Biol. 2009 Jul;41(7):1526–35.
9. Li X-L, Wong Y-S, Xu G, Chan JCN. Selenium-enriched Spirulina protects INS-1E
pancreatic beta cells from human islet amyloid polypeptide-induced apoptosis through
suppression of ROS-mediated mitochondrial dysfunction and PI3/AKT pathway. Eur J
Nutr. 2015 Jun;54(4):509–22.
10. Pithadia A, Brender JR, Fierke CA, Ramamoorthy A. Inhibition of IAPP Aggregation and
Toxicity by Natural Products and Derivatives. J Diabetes Res [Internet]. 2016 [cited 2016
May 1];2016. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4662995/
11. Quist A, Doudevski I, Lin H, Azimova R, Ng D, Frangione B, et al. Amyloid ion channels:
a common structural link for protein-misfolding disease. Proc Natl Acad Sci U S A. 2005
Jul 26;102(30):10427–32.
12. Mirzabekov TA, Lin MC, Kagan BL. Pore formation by the cytotoxic islet amyloid peptide
amylin. J Biol Chem. 1996 Jan 26;271(4):1988–92.
101

13. Green JD, Kreplak L, Goldsbury C, Li Blatter X, Stolz M, Cooper GS, et al. Atomic force
microscopy reveals defects within mica supported lipid bilayers induced by the
amyloidogenic human amylin peptide. J Mol Biol. 2004 Sep 17;342(3):877–87.
14. Engel MFM, Khemtémourian L, Kleijer CC, Meeldijk HJD, Jacobs J, Verkleij AJ, et al.
Membrane damage by human islet amyloid polypeptide through fibril growth at the
membrane. Proc Natl Acad Sci U S A. 2008 Apr 22;105(16):6033–8.
15. Lashuel HA. Membrane permeabilization: a common mechanism in protein-misfolding
diseases. Sci Aging Knowl Environ SAGE KE. 2005 Sep 21;2005(38):pe28.
16. Osman C, Voelker DR, Langer T. Making heads or tails of phospholipids in mitochondria.
J Cell Biol. 2011 Jan 10;192(1):7–16.
17. Kegulian NC, Sankhagowit S, Apostolidou M, Jayasinghe SA, Malmstadt N, Butler PC, et
al. Membrane Curvature-sensing and Curvature-inducing Activity of Islet Amyloid
Polypeptide and Its Implications for Membrane Disruption. J Biol Chem. 2015 Oct
23;290(43):25782–93.
18. Jayasinghe SA, Langen R. Lipid membranes modulate the structure of islet amyloid
polypeptide. Biochemistry (Mosc). 2005 Sep 13;44(36):12113–9.
19. Jayasinghe SA, Langen R. Membrane interaction of islet amyloid polypeptide. Biochim
Biophys Acta. 2007 Aug;1768(8):2002–9.
102

20. Luo P, Baldwin RL. Mechanism of helix induction by trifluoroethanol: a framework for
extrapolating the helix-forming properties of peptides from trifluoroethanol/water mixtures
back to water. Biochemistry (Mosc). 1997 Jul 8;36(27):8413–21.
21. Rohl CA, Baldwin RL. Comparison of NH exchange and circular dichroism as techniques
for measuring the parameters of the helix-coil transition in peptides. Biochemistry (Mosc).
1997 Jul 15;36(28):8435–42.
22. Apostolidou M, Jayasinghe SA, Langen R. Structure of alpha-helical membrane-bound
human islet amyloid polypeptide and its implications for membrane-mediated misfolding. J
Biol Chem. 2008 Jun 20;283(25):17205–10.
23. Hull RL, Westermark GT, Westermark P, Kahn SE. Islet amyloid: a critical entity in the
pathogenesis of type 2 diabetes. J Clin Endocrinol Metab. 2004 Aug;89(8):3629–43.
24. McMahon HT, Gallop JL. Membrane curvature and mechanisms of dynamic cell membrane
remodelling. Nature. 2005 Dec 1;438(7068):590–6.
25. Varkey J, Isas JM, Mizuno N, Jensen MB, Bhatia VK, Jao CC, et al. Membrane curvature
induction and tubulation are common features of synucleins and apolipoproteins. J Biol
Chem. 2010 Oct 15;285(42):32486–93.
26. Ardail D, Privat JP, Egret-Charlier M, Levrat C, Lerme F, Louisot P. Mitochondrial contact
sites. Lipid composition and dynamics. J Biol Chem. 1990 Nov 5;265(31):18797–802.
27. Horvath SE, Daum G. Lipids of mitochondria. Prog Lipid Res. 2013 Oct;52(4):590–614.
103

28. Gebert N, Joshi AS, Kutik S, Becker T, McKenzie M, Guan XL, et al. Mitochondrial
cardiolipin involved in outer membrane protein biogenesis: implications for Barth
syndrome. Curr Biol CB. 2009 Dec 29;19(24):2133–9.
29. Nielsen L, Khurana R, Coats A, Frokjaer S, Brange J, Vyas S, et al. Effect of environmental
factors on the kinetics of insulin fibril formation: elucidation of the molecular mechanism.
Biochemistry (Mosc). 2001 May 22;40(20):6036–46.

3.5 ACKNOWLEDGEMENTS
Research reported in this chapter was supported in whole or in part by the National Institutes on
Aging of the National Institutes of Health under award numbers AG027936 (RL).
The content of this manuscript is solely the responsibility of the authors and does not necessarily
represent the official views of the National Institutes of Health.

3.6 AUTHOR CONTRIBUTION STATEMENT
R.L. and A.K.O. contributed to the conception and design of the project; A.K.O. and K.T.
contributed to data acquisition, analysis and interpretation; A.K.O. and R.L. participated in drafts,
revisions and final approval of the manuscript.

104

3.7 COMPETING FINANCIAL INTERESTS STATEMENT
The authors have no competing financial interests to disclose.

3.8 ABBREVIATIONS
IAPP – islet amyloid polypeptide
T2DM – type 2 diabetes mellitus
CL – cardiolipin
PA – phosphatidic acid
POPC - 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine
PC – phosphatidylcholine
LUV – large unilamellar vesicle
CD – circular dichroism
TEM – electron microscopy
HFIP - Hexafluoro-2-propanol
ThT – Thioflavin T

3.9 METHODS
Materials. Thioflavin T (ThT) and Hexafluoro-2-propanol (HFIP) were obtained from Sigma-
Aldrich (St. Louis, MO). 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine (POPC) and
1',3'-bis[1,2-dioleoyl-sn-glycero-3-phospho]-sn-glycerol (cardiolipin) were obtained from Avanti
105

Polar Lipids (Alabaster, AL) as stocks concentrated in chloroform. Synthetic wild-type human
IAPP was obtained from Bachem (Torrance, CA).
Preparation of IAPP. IAPP was received as lyophilized powder and dissolved in HFIP to a
concentration of 0.67 mg/mL to disaggregate any aggregates and allow for aliquoting. Aliquots
were made in the HFIP and immediately frozen in N2 (l) and re-lyophilized for storage. Protein
concentrations were determined by UV absorbance at 280 nm in 8M guanidinium chloride using
an extinction coefficient of 1405 M
-1
cm
-1
and verified by CD spectroscopy upon resolublization.
Lyophilized IAPP stocks were stored, desiccated at room temperature in N 2 (g) desiccated under
a vacuum.
Preparation of Large Unilamellar Vesicles (LUV). Cardiolipin LUVs were made as in Chapter 2.
Briefly, we combined the indicated molar ratios of POPC with cardiolipin dissolved in chloroform,
then evaporated the chloroform in a borosilicate test tube to a dry film under a constant stream of
N2 (g) while spinning the tube to spread the lipid evenly while drying. These dried lipids were then
vacuum desiccated overnight in a desiccator flooded with N2 (g) prior to depressurization. Lipids
were reconstituted for use in 10 mM phosphate buffer pH 7.4 with or without NaCl and freeze-
thawed at least 6 times in an Eppendorf 1.5mL tube by passing back and forth between N2 (l) and
a hot water bath. Then lipids were passed through sequentially through a 400 and 100 nm a
polycarbonate filter 11 times each in a mini-extruder to homogenize their size to a diameter of
100nm.
Thioflavin T Fluorescence Assay. Thioflavin T (ThT) stocks were stored in water at -20°C at a
concentration of 5 mM and used at a concentration of 25μM as before (18). Lyophilized IAPP
samples solubilized in 10 mM phosphate buffer pH 7.4 with or without 100 mM NaCl were
106

transferred into a 2mm path length quartz cuvette. The solution was briefly measured in CD to
verify IAPP concentration and quality, then ThT was added and a baseline fluorescence reading
taken. CL-LUVs were then added and monitored in a JASCO FP-6500 spectrofluorometer with
the following conditions and settings: excitation wavelength = 450 nm, emission wavelength =
482 nm, excitation slit width = 1 nm, emission slit width = 10 nm, at room temperature.
t50 values were determined using math, as before (18). ThT emission intensities recorded over the
course of fibrilization were fitted to a sigmoidal curve based on the following equation (1)
(1) 𝐼 = ( 𝐼 𝑖 + 𝑚 𝑖 𝑡 )+
( 𝐼 𝑓 +𝑚 𝑓 𝑡 )
(1+𝑒 −(
𝑡 −𝑡 50
𝜏 )
)
,
where I = ThT fluorescence intensity, t = time, and t50 = time at half-maximal fluorescence
intensity. The other parameters are as described elsewhere (29). Each individual run was
normalized to the maximum intensity measured at the end of that experiment.
Circular Dichroism Spectroscopy. Lyophilized aliquots of IAPP were solubilized in 10mM
phosphate buffer pH 7.4 and transferred to a quartz cuvette (1 or 2 mm). CD spectra were taken in
a JASCO-815 spectropolarimeter from 190 to 260 nm. In cases where lipid concentration was
high, the farthest UV component of the spectra were too noisy for interpretation. In those cases,
the spectra were truncated to include only reliable data. Measurements were taken every 0.5 or 1
nm at a scan rate of 50 nm/min, with an averaging time of 1 s and background subtracted using
appropriate backgrounds.
Equations (2) and (3) were used to estimate the fraction of helicity based on established methods
(20)(21)
107

(2) 𝑓 𝛼𝐻
=
( 𝜃 𝑜𝑏 𝑠 −𝜃 𝑅𝐶
)
( 𝜃 𝐻 −𝜃 𝑅𝐶
)
,
where fαH is the fraction of helicity, θobs is the observed ellipticity, θRC is the value of ellipticity for
a peptide equal in length to IAPP and fully random coiled, and θH is the value of ellipticity for a
peptide equal in length to IAPP and fully helical. Mean residual ellipticity were converted from
raw ellipticity values taken at 222 nm using equation (3)
(3) 𝜃 =
𝜃 𝑟𝑎𝑤 𝑛𝐶𝑙
,
where θ is mean residual ellipticity (deg, cm
2
dmol
-1
), θraw is ellipticity in millidegrees, n is number
of amino acids, C is molar concentration, and l is path length in millimeters. This approach is can
be used to describe helix-coil transitions at constant temperature. Conditions for our experiments
were consistent with these criteria. As in our previous study, we used a value of For θH and θRC,
values of -34.7 x 10
3
deg cm
2
dmol
-1
and
-
1.5 x 10
3
deg cm
2
dmol
-1
were used to represent
completely helical and random coil structures, respectively (18). It remains a formal possibility
that, despite an excess of lipid, some IAPP remains free and thus, estimates of helicity should be
taken cautiously.
Electron Microscopy Studies. Sample were applied in 10 μL volumes to carbon-coated Formvar
films mounted on copper grids for ~10 minutes and excess liquid blotted on Whattman paper prior
to application of 1% uranyl acetate negative stain. After drying, a Jeol-1400 transmission electron
microscope operated at 100kV was used to image specimens.

108

Chapter 4
Chaperone-like Activity of Mitochondrially-Derived Peptides Captures
Diabetic Amyloid Seeds
Alan K. Okada
1*
, Kazuki Teranishi
1*
, Fleur Lobo
1
, Jialin Xiao
2
, Hemal H. Mehta
2
, J. Mario
Isas
1
, Kelvin Yen
2
, Pinchas Cohen
2
, Ralf Langen
1¥
1
Department of Biochemistry and Molecular Biology, Zilkha Neurogenetic Institute, University
of Southern California, Los Angeles, California 90033, USA,
2
University of Southern California
Davis School of Gerontology, Ethel Percy Andrus Gerontology Center, University of Southern
California, Los Angeles, CA 90089-0191, USA
¥
Corresponding Author: Ralf Langen, Zilkha Neurogenetic Institute, Department of
Biochemistry and Molecular Biology, University of Southern California, Los Angeles, California
90033, USA. Tel.: 323-442-1323 Fax: 323-442-4404. Email: Langen@usc.edu
* These authors contributed equally to this work

109

ABSTRACT
The mitochondrial-derived peptide (MDP) humanin (HN), and its S14G analog HNG have
emerged as wide spectrum stress response factors with strong cytoprotective effects. Early work
revealed this MDP to be protective in Alzheimer disease (AD) models. Recently, its value as a
therapeutic for diabetes has been reported. Many of its functions are attributed to its ability to
inhibit apoptosis, however, little is known about whether MDPs also have the capacity to facilitate
the cell’s unfolded protein response via direct interactions with amyloid proteins associated with
these diseases. Therefore, in this study, we have explored the effects of two MDPs, HNG and small
humanin-like peptide 2 (SHLP2), on the misfolding of islet amyloid polypeptide (IAPP), a critical
pathogenic step in type 2 diabetes mellitus (T2DM). A combination of fluorescence, electron
paramagnetic resonance spectroscopy, circular dichroism, and transmission electron microscopy,
indicates that both MDPs, HNG and SHLP2, inhibit the misfolding of IAPP by preferentially
binding to misfolded and seeding capable species of IAPP. These results suggest that HNG and
SHLP2 and likely many other MDPs have chaperone-like functions and thus carry the potential to
directly modify the disease process in T2DM. Along with direct anti-apoptotic activity and the
ability to beneficially modify metabolism, our data make humanin and SHLP2 exciting prospects
for the therapeutic treatment of T2DM and highlight the possibility that MDPs may represent a
family of mitochondrial stress response factors, many of which will be amenable to development
into therapeutics for protein-misfolding diseases.

110

4.1 INTRODUCTION
Mitochondrial-derived peptides, (MDPs) are a family of polypeptides encoded in distinct open
reading frames within the mitochondrial DNA (1–3). The 21 - 24 amino acid polypeptide, humanin
(HN), is the best-characterized among the identified MDPs (2). Single amino acid substitutions
have led to the discovery of humanin analogs with variable biological activity, such as the highly
potent S14G-HN (HNG) humanin analogue (4–7). HN is associated with improved longevity in
mouse models of aging (8). This, combined with a diverse set of biological features ranging from
cellular stress responses in multiple organ systems and tissues to modulation of metabolic activity
(1,9,10) have made humanin an attractive target for development as a therapeutic. Mirroring
humanin, other MDPs in the small humanin-like peptide (SHLP) family and MOTS-c have been
recently reported to display similarly diverse biological features, promoting viability and reducing
apoptosis in cell lines, as well as exhibiting beneficial insulin sensitizing effects, in vivo (2,3,11).
Thus, while humanin and its analogs are being investigated as potential therapeutics for
degenerative, protein-misfolding diseases such as age-related macular degeneration (AMD),
Alzheimer disease (AD), Creutzfeldt-Jakob disease (CJD) and diabetes mellitus (DM) (6,9,10,12–
15), it appears that multiple MDPs may have therapeutic potential. Consistent with a role for MDPs
as stress-inducible survival factors, the humanin homolog in rats, rattin, is upregulated in response
to stressful stimuli (16). In fact, humanin treatments rescue ROS-induced cytotoxicity in retinal
pigmented epithelial models of AMD (12), as well as amyloid-β (Aβ) (1) and prion-protein (PrP)
oligomer-induced toxicity in neurons (14), which are associated with AD and CJD, respectively.
Many of the activities attributed to HN and HNG are thought to be mediated via signaling through
cell surface receptors and interactions with intracellular apoptotic signaling molecules (17). Yet,
given the spectrum of protein-misfolding diseases in which HN is protective, surprisingly little is
111

known about whether humanin and other MDPs can directly influence the misfolding process via
chaperone-like mechanisms.
Here we directly investigate this notion using the 37-amino acid polypeptide, islet amyloid
polypeptide (IAPP), which plays a critical role in the pathogenesis of type 2 diabetes mellitus
(T2DM) (18). The misfolding and subsequent aggregation of IAPP induces a gain-of-function
associated with β-cell apoptosis, oxidative damage, mitochondrial dysfunction and ER stress (19–
22). In the islets of Langerhans, the process of IAPP misfolding and aggregation ultimately leads
to the replacement of pancreatic β-cell mass with deposits of fibrillar amyloid, which is the
hallmark of T2DM pathology (23). Within the cell, the formation of misfolded protein species puts
demands on the cell for molecular chaperones to mitigate their deleterious effects, however,
chaperone availability declines with age (24), as does MDP availability (2,10) and with mounting
stress, cellular proteostatic machinery can be overwhelmed leading to mitochondrial activation of
apoptotic processes (24). Given the paucity of information available regarding MDPs and protein-
misfolding, plus observations that humanin and its analogs delay diabetes onset and decrease β-
cell apoptosis (15), we sought to determine whether the potent humanin analog, HNG, as well as
the MDP, SHLP2, could function in a chaperone-like capacity to prevent the misfolding of IAPP.
We therefore used a combination of Thioflavin T (ThT) fluorescence studies in combination with
electron paramagnetic resonance spectroscopy (EPR), circular dichroism (CD) and transmission
electron microscopy (EM) to investigate the effect of MDPs, HNG and SHLP2, on IAPP
misfolding.

112

4.2 RESULTS
4.2.1 MDPs inhibit islet amyloid polypeptide misfolding.
In order to determine whether mitochondrially-derived peptides (MDPs) have the capacity to
inhibit the misfolding of islet amyloid polypeptide (IAPP), we first chose two candidate MDPs to
evaluate. We chose the HN analog HNG, where serine 14 is modified to a glycine residue, because
of its remarkably potent neuroprotective activity. Small humanin-like peptide 2 (SHLP2) was
chosen on account of its similar neuroprotective activity (2). We first monitored IAPP misfolding
in the presence of HNG or SHLP2 using thioflavin T. For this experiment the IAPP concentration
was held constant and the MDP concentrations varied. To avoid complications from potential
aggregation of HNG, we followed the peptide handling protocol first described by Arikawa et al.
2011 (25). Even so, we observed that prolonged storage (>1 month) of solubilized HNG stocks at
-20˚C markedly attenuated the potency of HNG and special care was taken to work with freshly
prepared HNG. With SHLP2 this effect was less pronounced, but similar care was taken to work
with fresh stocks of SHLP2. According to our ThT data, both MDPs inhibit IAPP aggregation
(Figure 4.1 a-c, Table 4.1). HNG exhibited the greater potency, essentially completely inhibiting
IAPP misfolding at substoichiometric concentrations, whereas closer to stoichiometric
concentrations of SHLP2 were required. The remarkable ability of HNG to perform at
substoichiometric concentrations (almost full inhibition at a molar ratio of 1:250, HNG:IAPP)
implies that it is unlikely for HNG to act on the bulk of the monomeric IAPP to inhibit aggregation.
Although binding to monomers could slow down aggregation by reducing the monomer pool
available for misfolding, the strongly substoichiometric ratios would only allow a small subset of
the IAPP molecules to be bound by HNG molecules. Such a minor reduction in available free
IAPP would have negligible effects on the misfolding kinetics. As SHLP2 inhibits near
113

stoichiometric concentrations, these data do not reveal whether SHLP2 inhibits misfolding via
interactions with the monomer or another target.
Figure 4.1

Figure 4.1 – Mitochondrial-derived peptides HNG and SHLP2 inhibit IAPP fibrilization by
thioflavin T. IAPP was allowed to misfold in 10mM potassium phosphate buffer, pH 7.4 at a
114

concentration of 12.5 μM in the presence of 25 μM ThT and monitored over 18 hours. MDP
concentrations ranged from 25 μM – 12.5 nM. A and B) Representative ThT kinetics traces of
IAPP misfolding in the presence of a) HNG or b) SHLP2. C) End point analysis of IAPP
misfolding by ThT fluorescence at 18 hours. Trend lines are shown for clarity. Concentrations of
MDPs used to inhibit IAPP misfolding are given in micromolar below the figure. Error bars
represent ± 1 standard deviation from a minimum of 3 experiments.
Table 4.1
IAPP t50
[MDP] μM HNG SHLP2
0 8.4 ± 1.3 8.4 ± 1.3
0.0125 13.4 ± 0.5 -
0.0167 12.4 ± 2.6 -
0.05 >20h -
0.25 >20h 10.2 ±2.7
1 >20h 11.1 ± 3.0
5 >20h 13.1 ± 2.3
7.14 - >20h
12.5 - >20h
25 >20h >20h

Table 4.1 – Kinetics analysis of IAPP misfolding monitored by ThT with various concentrations
HNG and e) SHLP2. t50 = time to half maximal fluorescence. >20h is given when no misfolding
115

is observed and a (-) is given when a condition was not tested. Values are given as a mean +/- 1
standard deviation from a minimum of 3 experiments.
4.2.2 EPR Spectroscopy reveals MDPs prevent loss of monomeric IAPP without
interactions with free IAPP
In order to better define the mechanisms by which MDPs inhibit IAPP misfolding, we used a
combination of site-directed spin labelling (SDSL) and continuous wave-electron paramagnetic
resonance (EPR) spectroscopy. EPR spectra are highly sensitive to the mobility of the spin label
and changes in their amplitudes can be used to monitor binding interactions and aggregation of
free, monomeric IAPP. Stoichiometric equivalents of MDPs were used for this experiment so that
any MDP-IAPP(monomer) interactions would yield robust changes in central line amplitudes.
Spectra were recorded for IAPP spin-labelled at amino acid 33 (IAPP33R1) and 17 (IAPP17R1)
(26) in the presence or absence of MDPs and their amplitudes were plotted (Figure 4.2 a-d). Spectra
for both IAPP33R1 and IAPP17R1 showed sharp lines characteristic of a predominantly
monomeric structure. When IAPP33R1 or IAPP17R1 were solubilized in the presence of HNG or
SHLP2, we observed remarkably similar spectra for IAPP alone and in the presence of either MDP.
Thus, no direct binding interactions between IAPP monomers and MDPs could be detected under
the conditions used. Such binding interactions would have reduced the overall tumbling of IAPP
and thereby yielded broader lines. In addition, rigidification by binding or structuring within ~10
amino acids of the spin-labeled sidechains would have further resulted in line broadening (27).
The lack of either of these effects is consistent with our hypothesis that HNG performs its essential
inhibitory function by binding to and stabilizing something other than the monomer. It also implies
that SHLP2 functions similarly, despite having a markedly lower potency.
116

By following amplitude changes of IAPP33R1 EPR spectra over time, we can monitor misfolding
in the presence and absence of MDPs as this region becomes tightly packed in a parallel in-register
structure upon misfolding (26). When compared to untreated controls, both MDPs inhibited
misfolding. In fact, we observed a remarkable retention of central line amplitude over the course
of the experiment, with a slight but steady loss of amplitude in the case of both MDP treatments
(Figure 4.2 e, f). These data demonstrate that the presence of MDPs help the naïve population of
IAPP largely retains its free, unaggregated state in solution. The slight changes in EPR signal
indicate that, despite the presence of MDPs, small amounts of higher order IAPP species can still
form.
117

Figure 4.2

Figure 4.2 – CW-EPR spectroscopy shows MDPs prevent loss of IAPP monomer without
detectable binding to free IAPP. IAPP33R1 or IAPP17R1 was incubated at room temperature in
10mM phosphate buffer, pH 7.4 in the presence or absence of stoichiometric equivalents of MDPs.
118

A - D) EPR spectral overlay of a) IAPP33R1 alone (black) and with HNG (red), b) IAPP33R1
alone (black) and with SHLP2 (red), c) IAPP17R1 alone (black) and with HNG (red), d)
IAPP17R1 alone (black) and with SHLP2 (red). E) IAPP misfolding kinetics monitored over the
course of fibrilization by central line amplitude. Symbols represent the following: (□) IAPP33R1,
(◊) IAPP33R1 + HNG, (∆) IAPP33R1 + SHLP2. Error bars represent +/- 1 standard deviation. p-
values of MDP treated vs. untreated IAPP33R1 at 1 hr <0.05, and at 2+ hr <0.01. F) Average
central line amplitudes for MDP treated IAPP33R1 at hour 10, * p = 0.001. Error bars represent
+/- 1 standard deviation from a minimum of 3 experiments.

4.2.3 CD spectroscopy and TEM indicates MDPs prevent the misfolding of IAPP
In order to monitor the secondary structure changes during IAPP misfolding in the presence or
absence of MDPs, we performed time-resolved circular dichroism (CD). For these experiments we
mimicked the conditions used in the EPR experiments, that is, 10 mM phosphate buffer, pH 7.4
with stoichiometric equivalents of IAPP and HNG or SHLP2. We first monitored secondary
structural changes of IAPP and MDPs taken in isolation. As expected, IAPP alone in solution is
initially largely disordered as indicated by a negative peak at ~202 nm (Figure 4.3 a, d). Over the
course of fibrilization, IAPP develops a negative peak at 218 nm, indicating the formation of β-
sheet rich conformers. Electron microscopy confirms the development of IAPP fibrils (Figure 4.3
g). On the other hand, HNG and SHLP2 in the same conditions retain a primarily random coil
signal throughout the duration of the experiment (Supplemental Figure 4.S1 a-d).
We then used CD to monitor IAPP misfolding in the presence of MDPs. In this case, we find that
the spectra of each peptide measured in isolation add, arithmetically, to closely match spectra taken
119

of the peptides together, showing largely disordered conformations (Figure 4.S2 a, b). These data
agree with our EPR findings showing no significant interactions between the free IAPP population
and either MDP.
We next used CD to follow IAPP treated with MDPs over time. Whether treated with HNG or
SHLP2, the IAPP MDP mixtures mostly retained their starting structures, predominantly random
coil, (Figure 4.3 b, c, e, f). The notion that we are observing largely monomeric samples devoid of
significant aggregates is consistent with EM images of MDP treated IAPP (Figure 4.3 h, i). Here
we see small structures approaching the limit of our resolution in MDP treated samples of IAPP
that may represent multimeric MDP complexes or IAPP:MDP complexes. The CD and EM data
are in direct agreement with our EPR studies and validate the notion that MDPs maintain the bulk
of IAPP in a monomeric state.
Reminiscent of our EPR observations, there is a small loss of random coil signal over time in IAPP
samples treated with either HNG or SHLP2 (Figure 4.3 e, f, dashed lines). While the CD data do
not indicate which of the peptides is responsible for this slow conformational change, our EPR
data indicate that structural changes in the IAPP must have at least in part been responsible for the
observed changes seen by CD.
120

Figure 4.3

Figure 4.3 – Time-resolved circular dichroism and transmission electron microscopy confirm
MDPs prevent loss of IAPP monomers and prevent fibrilization. A-C) Circular dichroism spectra
of 15μM IAPP treated with molar equivalents of MDPs recorded at the beginning and end of each
experiment. Where IAPP and MDPs are measured together, data are displayed as a weighted mean
121

residual ellipticity (MRE`). Details regarding the calculation of MRE` are given in the Materials
and Methods section. D-F) Time resolved CD of experiments in (a) - (c). Ellipticities were
recorded at 202 nm and 218 nm to follow transitions from random coil (202 nm) to β-sheet (218
nm). Traces represent an average of at least 3 experiments. G-I) Electron micrographs taken at the
end of the experiment for G) IAPP alone, H) IAPP treated with HNG, and I) IAPP treated with
SHLP2. Scale bars equal 200 nm.
Figure 4.S1

122

Figure 4.S1 – HNG and SHLP2 structural stability during time-resolved circular dichroism. A)
Spectra of HNG taken at the beginning and end of time-resolved CD experiments. B) Spectra of
SHLP2 at the beginning and end of time-resolved CD experiments. C) Time-resolved ellipticity of
HNG recorded at 202 and 218 nm. D) Time-resolved ellipticity of SHLP2 recorded at 202 and 218
nm. Traces are an average of at least 3 experiments.
Figure 4.S2

Figure 4.S2 – IAPP and MDPs alone and together remain primarily disordered in solution. A)
CD spectra of the arithmetic sum of IAPP and HNG measured alone, [IAPP] + [HNG] (solid line),
overlaid with the spectra of IAPP and HNG mixed in solution, [IAPP + HNG] (dotted line). B)
Spectra of IAPP and SHLP2 measured separately, [IAPP] + [SHLP2] (solid line), overlaid with
IAPP and SHLP2 together in solution, [IAPP + SHLP2] (dotted line). Traces are an average of at
least 3 experiments.

123

Figure 4.S3

Figure 4.S3 – HNG and SHLP2 form small oligomeric structures. A) HNG and B) SHLP2 were
visualized by transmission electron microscopy and reveal small darkly staining structures.
4.2.4 MDPs, HNG and SHLP2, prevent propagation by IAPP seeds
The ThT, EPR and CD are inconsistent with a mechanism in which the MDPs affect misfolding
via binding to IAPP monomers. An alternative way in which MDPs could prevent misfolding
without interacting with the monomer is to prevent seeds from promoting misfolding. In order to
test this, we sonicated pre-formed IAPP fibrils and used them to seed freshly dissolved IAPP. Once
again we monitored the misfolding of IAPP in the presence or absence of HNG or SHLP2 using
ThT. As is expected from a seeded reaction, we observed that seeding bypasses the lag phase in
the absence of MDPs. Furthermore, we found that both HNG and SHLP2 displayed remarkable
capacity to inhibit the seeded IAPP reaction (Figure 4.4 a, b). This, in concert with our earlier ThT,
EPR and CD data implies that MDPs bind directly to misfolded IAPP seeds.
124

Figure 4.4

Figure 4.4 – MDPs protect against seeding of pre-formed IAPP aggregates. A) Representative
fluorescence kinetics traces from 2.5 mol % seeded 12.5 μM IAPP reactions treated with HNG or
SHLP2. B) Bar graph of averaged and normalized fluorescence data from (a) at 10 hours. Error
bars represent +/- standard deviation from at least 3 independent measurements. * p < 0.01.
4.2.5 MDPs bind directly to misfolded IAPP seeds
In order to test for a direct interaction between the MDPs and IAPP seeds, we employed a co-
sedimentation assay. Sonicated seeds, despite their reduced size, can be pelleted through
ultracentrifugation, allowing us to enrich pellets with IAPP seeds and anything that they pull down
with them. We, therefore, incubated sonicated pre-formed IAPP fibrils with HNG or SHLP2 and
pelleted the seed fraction. Membranes dotted with supernatant and pellet samples were probed
with antibodies against either HNG (anti-HN AP) or SHLP2 (anti-SHLP2). In both cases, we found
that the pelleted fraction was enriched in MDPs (Figure 4.5), verifying that HNG and SHLP2
125

directly interact with IAPP seeds. This finding is consistent with a mechanism wherein MDPs
inhibit IAPP misfolding by preferentially targeting misfolded IAPP. As small volumes of
supernatant were allowed to remain in the pellet fraction to prevent accidental removal of the
pellet, it is expected that trace amounts of soluble MDP are observed in the pellet fractions from
samples containing only HNG or SHLP2.
Figure 4.5

Figure 4.5 – Co-sedimentation assay shows HNG and SHLP2 bind IAPP seeds. A) Representative
dot blot from supernatant (top) and pelleted (bottom) fractions of HNG and sonicated pre-formed
IAPP fibrils (HNG + IAPP), HNG alone (HNG), or sonicated pre-formed IAPP fibrils alone
(IAPP), probed with anti-HN AP antibody. B) Representative dot blot from supernatant (top) and
pelleted (bottom) fractions of SHLP2 and sonicated pre-formed IAPP fibrils (SHLP2 + IAPP),
SHLP2 alone (SHLP2), or sonicated pre-formed IAPP fibrils alone (IAPP), probed with anti-
SHLP2 antibody.

126

4.3 DISCUSSION
Here we show that the mitochondrial-derived peptides HNG and SHLP2 prevent the misfolding
of islet amyloid polypeptide. Multiple lines of evidence indicate that the mechanism underlying
this activity is the binding of misfolded IAPP seeds by the two MDPs. The strongly sub
stoichiometric inhibition in the case of HNG rules out the necessity for bulk capture of monomeric
IAPP by HNG. Moreover, the EPR data do not detect any binding interaction between IAPP and
either MDP, while the CD data reveal that mixing of IAPP and MDPs does not result in any
detectable changes in secondary structure. These data are inconsistent with any substantive binding
of the free pool of naïve IAPP by either MDP. On the other hand, we observe direct interactions
between the MDPs and IAPP misfolding seeds by co-sedimentation. This interaction is sufficient
to robustly prevent seeded misfolding reactions indicating that these MDPs have the capacity to
prevent such seeds from functioning as a template for the misfolding of naïve IAPP. Consistent
with this notion, our CD and EPR data further demonstrate that the hallmark cooperativity
normally observed in IAPP misfolding reactions can be blocked by MDP treatments.
A subtle feature of the MDP inhibition of IAPP misfolding is the slow decrease in free IAPP signal
observed in CD and EPR studies. Combined with the lack of cooperativity in the process of
oligomerization, these data suggest that some higher order species can, in fact form, in the presence
of MDPs but that they cannot promote seeding. In light of these findings, it seems unlikely that
MDPs act by completely blocking the formation of all oligomeric and fibrillar species, including
seeds. Rather, these data are more consistent with a mechanism wherein MDPs bind to any seeds
that may have formed, and, in doing so, inhibit misfolding. Such a mechanism is akin to the “cap
and contain” mechanism previously described for the chaperonins TRiC and CCT5, which bind to
misfolded huntingtin in a manner that blocks seeding-mediated misfolding (28). A potential heat
127

shock protein-like function of MDPs, as a family, may also explain why prior studies found HNG
to inhibit Aβ aggregation (29). Future work will be required to precisely resolve the mechanism
by which MDPs inhibit IAPP seeding, yet the heat shock-like functionality, when taken in context
with the known neuroprotective effects of humanin in the setting of Aβ toxicity (AD pathology)
raises the possibility that humanin and other MDPs serve a role as a sensor or a response to amyloid
misfolding. If true, this would expand the function of mitochondria in amyloid disease beyond its
apoptotic role to that of an early warning component of the cellular defense system.
These data also highlight the therapeutic potential of HNG and SHLP2 in the treatment of T2DM.
Our data reveal the interactions of MDPs with IAPP to have a degree of specificity for non-
monomeric IAPP species. This quality is likely to help reduce unintended interactions that might
prevent IAPP from performing its regular physiological function, but instead target only dangerous
conformations. Taken in context with the pleiotropic qualities already assigned to humanin, these
considerations make therapeutic approaches with humanin and perhaps other MDPs like SHLP2,
whether via genetic upregulation of endogenous MDPs or through exogenous administration of
MDPs and their analogues, an exciting avenue for further research. Furthermore, it will be
important to investigate whether other MDPs have similar cytoprotective qualities against other
forms of protein misfolding diseases.

128

4.4 REFERENCES
1. Yen K, Lee C, Mehta H, Cohen P. The emerging role of the mitochondrial-derived peptide
humanin in stress resistance. J Mol Endocrinol. 2013 Feb;50(1):R11-19.
2. Cobb LJ, Lee C, Xiao J, Yen K, Wong RG, Nakamura HK, et al. Naturally occurring
mitochondrial-derived peptides are age-dependent regulators of apoptosis, insulin
sensitivity, and inflammatory markers. Aging. 2016 Apr 10;
3. Lee C, Zeng J, Drew BG, Sallam T, Martin-Montalvo A, Wan J, et al. The mitochondrial-
derived peptide MOTS-c promotes metabolic homeostasis and reduces obesity and insulin
resistance. Cell Metab. 2015 Mar 3;21(3):443–54.
4. Yamagishi Y, Hashimoto Y, Niikura T, Nishimoto I. Identification of essential amino acids
in Humanin, a neuroprotective factor against Alzheimer’s disease-relevant insults. Peptides.
2003 Apr;24(4):585–95.
5. Terashita K, Hashimoto Y, Niikura T, Tajima H, Yamagishi Y, Ishizaka M, et al. Two serine
residues distinctly regulate the rescue function of Humanin, an inhibiting factor of
Alzheimer’s disease-related neurotoxicity: functional potentiation by isomerization and
dimerization. J Neurochem. 2003 Jun;85(6):1521–38.
6. Hashimoto Y, Niikura T, Tajima H, Yasukawa T, Sudo H, Ito Y, et al. A rescue factor
abolishing neuronal cell death by a wide spectrum of familial Alzheimer’s disease genes
and Abeta. Proc Natl Acad Sci U S A. 2001 May 22;98(11):6336–41.
129

7. Hashimoto Y, Niikura T, Ito Y, Sudo H, Hata M, Arakawa E, et al. Detailed characterization
of neuroprotection by a rescue factor humanin against various Alzheimer’s disease-relevant
insults. J Neurosci Off J Soc Neurosci. 2001 Dec 1;21(23):9235–45.
8. Lee C, Wan J, Miyazaki B, Fang Y, Guevara-Aguirre J, Yen K, et al. IGF-I regulates the
age-dependent signaling peptide humanin. Aging Cell. 2014 Oct;13(5):958–61.
9. Kuliawat R, Klein L, Gong Z, Nicoletta-Gentile M, Nemkal A, Cui L, et al. Potent humanin
analog increases glucose-stimulated insulin secretion through enhanced metabolism in the
β cell. FASEB J Off Publ Fed Am Soc Exp Biol. 2013 Dec;27(12):4890–8.
10. Muzumdar RH, Huffman DM, Atzmon G, Buettner C, Cobb LJ, Fishman S, et al. Humanin:
a novel central regulator of peripheral insulin action. PloS One. 2009;4(7):e6334.
11. Fuku N, Pareja-Galeano H, Zempo H, Alis R, Arai Y, Lucia A, et al. The mitochondrial-
derived peptide MOTS-c: a player in exceptional longevity? Aging Cell. 2015
Dec;14(6):921–3.
12. Sreekumar PG, Ishikawa K, Spee C, Mehta HH, Wan J, Yen K, et al. The Mitochondrial-
Derived Peptide Humanin Protects RPE Cells From Oxidative Stress, Senescence, and
Mitochondrial Dysfunction. Invest Ophthalmol Vis Sci. 2016 Mar 1;57(3):1238–53.
13. Niikura T, Chiba T, Aiso S, Matsuoka M, Nishimoto I. Humanin: after the discovery. Mol
Neurobiol. 2004 Dec;30(3):327–40.
14. Sponne I, Fifre A, Koziel V, Kriem B, Oster T, Pillot T. Humanin rescues cortical neurons
from prion-peptide-induced apoptosis. Mol Cell Neurosci. 2004 Jan;25(1):95–102.
130

15. Hoang PT, Park P, Cobb LJ, Paharkova-Vatchkova V, Hakimi M, Cohen P, et al. The
neurosurvival factor Humanin inhibits beta-cell apoptosis via signal transducer and activator
of transcription 3 activation and delays and ameliorates diabetes in nonobese diabetic mice.
Metabolism. 2010 Mar;59(3):343–9.
16. Jia Y, Lue Y-H, Swerdloff R, Lee K-W, Cobb LJ, Cohen P, et al. The cytoprotective peptide
humanin is induced and neutralizes Bax after pro-apoptotic stress in the rat testis.
Andrology. 2013 Jul;1(4):651–9.
17. Lee C, Yen K, Cohen P. Humanin: a harbinger of mitochondrial-derived peptides? Trends
Endocrinol Metab TEM. 2013 May;24(5):222–8.
18. Hull RL, Westermark GT, Westermark P, Kahn SE. Islet amyloid: a critical entity in the
pathogenesis of type 2 diabetes. J Clin Endocrinol Metab. 2004 Aug;89(8):3629–43.
19. Haataja L, Gurlo T, Huang CJ, Butler PC. Islet amyloid in type 2 diabetes, and the toxic
oligomer hypothesis. Endocr Rev. 2008 May;29(3):303–16.
20. Konarkowska B, Aitken JF, Kistler J, Zhang S, Cooper GJS. Thiol reducing compounds
prevent human amylin-evoked cytotoxicity. FEBS J. 2005 Oct;272(19):4949–59.
21. Ma ZA, Zhao Z, Turk J. Mitochondrial Dysfunction and β-Cell Failure in Type 2 Diabetes
Mellitus. Exp Diabetes Res [Internet]. 2012 [cited 2016 May 4];2012. Available from:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3216264/
22. Huang C, Lin C, Haataja L, Gurlo T, Butler AE, Rizza RA, et al. High expression rates of
human islet amyloid polypeptide induce endoplasmic reticulum stress mediated beta-cell
131

apoptosis, a characteristic of humans with type 2 but not type 1 diabetes. Diabetes. 2007
Aug;56(8):2016–27.
23. Alejandro EU, Gregg B, Blandino-Rosano M, Cras-Méneur C, Bernal-Mizrachi E. Natural
history of β-cell adaptation and failure in type 2 diabetes. Mol Aspects Med. 2015
Apr;42:19–41.
24. Hipp MS, Park S-H, Hartl FU. Proteostasis impairment in protein-misfolding and -
aggregation diseases. Trends Cell Biol. 2014 Sep;24(9):506–14.
25. Arakawa T, Niikura T, Kita Y. The biological activity of Humanin analogs correlates with
structure stabilities in solution. Int J Biol Macromol. 2011 Jul 1;49(1):93–7.
26. Bedrood S, Li Y, Isas JM, Hegde BG, Baxa U, Haworth IS, et al. Fibril Structure of Human
Islet Amyloid Polypeptide. J Biol Chem. 2012 Feb 17;287(8):5235–41.
27. Margittai M, Fasshauer D, Jahn R, Langen R. The Habc domain and the SNARE core
complex are connected by a highly flexible linker. Biochemistry (Mosc). 2003 Apr
15;42(14):4009–14.
28. Darrow MC, Sergeeva OA, Isas JM, Galaz-Montoya J, King JA, Langen R, et al. Structural
Mechanisms of Mutant Huntingtin Aggregation Suppression by Synthetic Chaperonin-like
CCT5 Complex Explained by Cryo-electron Tomography. J Biol Chem. 2015 May
20;jbc.M115.655373.
132

29. Zhang W, Du Y, Bai M, Xi Y, Li Z, Miao J. S14G-humanin inhibits Aβ1-42 fibril formation,
disaggregates preformed fibrils, and protects against Aβ-induced cytotoxicity in vitro. J Pept
Sci Off Publ Eur Pept Soc. 2013 Mar;19(3):159–65.
30. Jayasinghe SA, Langen R. Lipid membranes modulate the structure of islet amyloid
polypeptide. Biochemistry (Mosc). 2005 Sep 13;44(36):12113–9.

4.5 ACKNOWLEDGEMENTS
The project described was supported by the Diabetes & Obesity Research Institute (DORI) with
funding from the Stewart Clifton Endowment. The content is solely the responsibility of the
authors and does not necessarily represent the official views of the DORI or the Stewart Clifton
Endowment.

4.6 AUTHOR CONTRIBUTION STATEMENT
A.K.O., K.T., J.X., H.H.M., K.Y., P.C., and R.L., contributed to the conception and design of the
project; A.K.O., K.T., F.L., J.X., and J.M.I., contributed to data acquisition, analysis and
interpretation; A.K.O., K.T., F.L., J.X., H.H.M., J.M.I., K.Y., P.C., and R.L. participated in drafts,
revisions and final approval of the manuscript being submitted for publication.

4.7 COMPETING FINANCIAL INTERESTS STATEMENT
P.C. is a consultant and stockholder of CohBar Inc.
133

4.8 ABBREVIATIONS
HN – Humanin
HNG – Humanin S14G
SHLP – Small humanin-like peptide
MDP – Mitochondrial-derived peptide
IAPP – islet amyloid polypeptide
T2DM – type 2 diabetes mellitus
ThT – Thioflavin T
EPR – Electron paramagnetic resonance
CD – circular dichroism
TEM – electron microscopy
HFIP – Hexafluoro-2-propanol

4.9 METHODS
Materials and chemicals - Wild-type human IAPP was purchased from Bachem Bioscience Inc.
(King of Prussia, PA). Hexafluoroisopropanol was purchased from Sigma-Aldrich. 1-oxyl-2,2,5,5-
tetramethyl-Δ3-pyrroline-3-methyl methanethiosulfonate (MTSL), was purchased from Toronto
Research Chemicals (Toronto, Ontario, Canada). Human IAPP cysteine mutants with alanine
substitutions for the native cysteines at positions 2 and 7 were purchased from Biomer Technology
(Pleasanton, CA). Humanin S14G (HNG) was obtained from Genscript (Piscataway, NJ). Small
humanin-like peptide 2 (SHLP2) was obtained from CPC Scientific (San Jose, CA). SuperBlock
T20(PBS) Blocking Buffer was obtained from Thermo Scientific (Rockford, IL).
134

Peptide handling and storage – Lyophilized wild type human IAPP was dissolved in HFIP,
aliquoted into individual tubes and flash frozen in N2 (l) prior to lyophilization. UV absorbance at
280 nm was used to determine IAPP protein concentrations in denaturing conditions (8M
guanidinium chloride) using an ε280 of 1405 M
-1
cm
-1
and verified by CD spectroscopy upon
resolublization. Lyophilized IAPP stocks were stored in N2 (g) under vacuum. MTSL labeled IAPP
was stored at
-
20˚C in HFIP. HNG was obtained lyophilized from Genscript and stored lyophilized
at
-
80˚C until solubilization. HNG was solubilized in water at 1 mg/mL and aliquoted. Aliquots
were stored at
-
20˚C until use. SHLP2 was obtained lyophilized and stored at
-
80˚C until used.
SHLP2 was solubilized in water at 1 mg/mL, aliquoted and stored at
-
20˚C until use.
Peptide labeling –Spin labelling was performed as before (26). Briefly, Single cysteine mutants of
IAPP with MTSL (>5 molar excess) for ~1 h at room temperature. Excess MTSL was removed
via cation exchange using a Toyopearl cation exchange column and subsequently desalted on a
C18 reverse phase SpinColumn (Harvard Apparatus, Holliston, MA), and ultimately eluted in
HFIP. Spin labelled peptide was stored in HFIP at -80˚C. Peptide concentration was verified at the
beginning of each experiment by comparing central line amplitudes and double integral values
against a standard concentration curve on the EPR apparatus.
Thioflavin T Fluorescence Studies – Thioflavin T (ThT) was stored at a 5 mM stock concentration
in water at -20°C. ThT was used at a 25μM final concentration to monitor IAPP misfolding. IAPP
aliquots were prepared as above. Individual samples of IAPP were solubilized in appropriate buffer
with ThT to a concentration of 12.5μM from a dry powder. MDPs were prepared to a stock
concentration of 1mg/mL as described above and added to appropriate reactions. The mixtures
were monitored for fluorescence in a 2 mm quartz cuvette and a JASCO FP-6500
spectrofluorometer at room temperature. Fluorescence was monitored under the following settings
135

and conditions: excitation wavelength = 450 nm, emission wavelength = 482 nm, excitation slit
width = 1 nm, emission slit width = 10 nm, and pH = 7.4. t50 values were determined as before
(30), using a sigmoidal model to fit our data. Each experiment was normalized to an appropriate
IAPP control by dividing fluorescence intensities by the maximal IAPP control intensity.
For seeding experiments IAPP was fibrilized in 10 mM phosphate buffer, pH 7.4 for 2 weeks at
55 μM. Fibrils were sonicated using a titanium tip sonicator 4 x 30 seconds each and placed on ice
in between sonications. ThT, IAPP and MDPs were prepared as described above. Fluorescence
was measured in an Eppendorf AF2200 multi-well fluorescence plate reader. Reaction volumes
were ~ 100 μL. Fluorescence was monitored under the following settings and conditions:
excitation wavelength = 450 nm, emission wavelength = 482 nm, 25 flashes per measurement and
a gain setting of 75.
Co-sedimentation Assay – For determination of direct interaction between HNG and IAPP, 4:1
wt/wt ratios of HNG and sonicated pre-formed IAPP fibrils (1µg:250ng, respectively), or 1:1 wt/wt
ratios of SHLP2 and sonicated pre-formed fibrils (1µg:1µg, respectively) were co-incubated in a
50 µL reaction volume and centrifuged at 55,000 rpm for 30 minutes at 4 °C. 90% of the reaction
volumes were removed and labeled as supernatant. 20 µL of buffer was added to the pellet fraction
to facilitate efficient recovery. Buffer was added to supernatant fractions to appropriately match
pellet dilution and 2 µL dots from supernatant and pellet samples were placed on nitrocellulose
membranes then dried. Membranes were subsequently blocked for 30 - 60 minutes in either milk
or SuperBlock T20(PBS) Blocking Buffer and probed with antibody against either HNG (1:1000,
α-HN AP; chicken) or SHLP2 (1:500, α-SHLP2; rabbit) for 1 hour at room temperature or
overnight at 4 °C. Primary antibody was recovered and membranes were washed 3 x 5 min in tris
buffered saline with 0.005% tween 20 (TBST). HNG and SHLP2 treated membranes were then
136

probed with IRdye 800 donkey α-chicken (1:10,000 Li-Cor) and IRdye 680 donkey α-rabbit
(1:10,000 Li-Cor). Membranes were washed 3 x 15 minutes in TBST and imaged on a Li-Cor
Odyssey fluorescent imager.
Circular Dichroism Spectroscopy – IAPP was prepared as above. CD spectra were measured
between 195 and 260 nm in a Jasco-815 spectropolarimeter. Spectra were scanned at a rate of 50
nm/min and measurements taken every 0.5 nm, with an averaging time of 1 second and background
subtracted using appropriate backgrounds. Time-resolved measurements were taken at 202 or 218
nm and averaged over 20 seconds. Ellipticity values were converted into mean residual ellipticity
using equation (1)
(1) 𝜃 =
𝜃𝑟𝑎𝑤 𝑛𝐶𝑙
,
where θ is the mean residual ellipticity expressed in degrees cm
2
dmol
-1
, θraw is the measured
ellipticity in millidegrees, n is the number of amino acids, C is the molar concentration, and l is
the path length in millimeters.
Mean residual ellipticity calculations in mixed CD samples (MRE`) – For CD experiments where
IAPP and either HNG or SHLP2 were measured together, MRE` values are sometimes reported.
We calculated MRE` values by converting equation (1) to equation (2)
(2) 𝜃 =
𝜃 𝑟𝑎𝑤 ( ( 𝑛 1
×𝐶 1
) +( 𝑛 2
×𝐶 2
) ) 𝑙 ,
Where n1 and C1 represent the number of amino acids and concentration of one peptide,
respectively, and n2 and C2 represent the number of amino acids and concentration of the second
peptide.
137

For initial interaction studies, addition of IAPP and MDP alone values were performed using
JASCO Spectra Manager 2.0 software. In this case, ellipticities are reported in their raw form.
Electron Paramagnetic Resonance Spectroscopy– Cys-less IAPP33C was labeled with MTSL
(IAPP33R1) and stored as described above. Stock IAPP33R1 was dried in a constant stream of N2
(g) and solubilized to a concentration of ~15 μM in 10 mM phosphate buffer pH 7.4 with either
vehicle or stoichiometric equivalents of MDP. Samples were drawn up into glass capillaries (0.6
mm diameter, VitroCom, Mountain Lakes, NJ) and sealed at the end. Spectra were recorded on a
Bruker EMX spectrometer (Billerica, MA) at 12 mW with a magnetic field scan range of 100
Gauss in an HS resonator. Measurements of central line amplitude were made by taking the
difference between the zenith and nadir of the central line. Values were normalized to the starting
amplitude.
Transmission Electron Microscopy – Samples were applied to carbon-coated Formvar films
mounted on copper grids for at least 10 minutes and negative stained with 1% uranyl acetate after
excess liquid was blotted away. Grids were imaged on a Jeol-1400 transmission electron
microscope operated at 100kV.

138

Appendix A
Lysine Acetylation is a Mechanism for Regulating the Interaction between Proteins and
Membranes

Alan K. Okada
1*
, Mark R. Ambroso
1*
, Joo-Yeun Lee
2
, Arthur Alves Melo
3
, Daniel Merken
4
,
Kazuki Teranishi
1
, Oliver Daumke
3
, Karen Chang
2
, Ian S. Haworth
4
, and Ralf Langen
1,¥

1
Zilkha Neurogenetic Institute, Department of Biochemistry and Molecular Biology, University
of Southern California, Los Angeles, California 90033, USA,
2
Zilkha Neurogenetic Institute and
Department of Cell & Neurobiology, Keck School of Medicine, University of Southern
California, Los Angeles, California, United States of America ; Neuroscience Graduate Program,
University of Southern California, Los Angeles, California, United States of America
3
Max-
Delbrück-Center for Molecular Medicine, Crystallography, Robert-Rössle-Straße 10, 13092
Berlin, Germany; Institute of Chemistry and Biochemistry, Freie Universität Berlin, Takustraße
6, 14195 Berlin, Germany; Institute of Medical Physics and Biophysics, Charité, Charitéplatz 1,
10117 Berlin, Germany,
4
Department of Pharmacology and Pharmaceutical Sciences, University
of Southern California, Los Angeles, California 90089, USA,
¥
Corresponding Author: Ralf Langen, Tel.: 323-442-1323 Fax: 323-442-4404. Email:
Langen@usc.edu
*These authors contributed equally to this research.

139

A.1 SUMMARY
Post-translational modifications regulate protein function in vivo. Acetylation of lysine residues is
well-known to regulate protein-DNA (1,2) and protein-protein (2)

interactions by neutralizing
lysine positive charges. Membrane-binding proteins often utilize positively charged lysines to
interact with negatively charged regions of the membrane (3). Such membrane interactions can
perform a myriad of cellular functions, yet whether they are regulated by lysine acetylation remains
largely unexplored. Here we developed a bioinformatics approach that correlates existing post-
translational modification and structural database information. From an analysis of 48 proteins, we
find that lysine acetylation is predominantly localized to membrane-binding lysines of
bin/amphiphysin/rvs (BAR), phox-homology (PX), C2, and Eps15-homology (EHD) lipid-binding
domains when compared to residues outside of the protein-membrane interface. Acetylation-
mimicking mutants of BAR, EHD, and C2 domain containing proteins (amphiphysin, EHD2, and
synaptotagmin 1) altered their remodeling of liposomes and/or reduced binding. When expressed
in cell culture, acetylation-mimicking mutants of amphiphysin and EHD2 also robustly altered
membrane-binding and subcellular localization, favoring cytosolic over membrane-associated
distributions. Transgenic flies harboring amphiphysin acetylation-mimicking mutations displayed
severely disrupted T-tubule organization that led to a flightless phenotype. Taken together, our
bioinformatic, biophysical, cellular and animal studies show that lysine acetylation preferentially
occurs widely within membrane-binding regions of proteins where it potently regulates membrane
binding and function. This work expands the known role of acetylation beyond control of protein-
protein and protein-DNA interactions to include protein-membrane interactions. It further adds a
level of complexity to the regulation of membrane functions as the combination of acetylation with
phosphorylation and other PTMs creates multiple potential PTM-driven functional modes.
140

Furthermore, our bioinformatics approach for studying structural details of large cohorts of
membrane-binding proteins can be used to correlate structural features of proteins with any
posttranslational modification.

141

A.2 RESULTS
A.2.1 Bioinformatics analysis reveals functional specificity for lysine acetylation at
the membrane binding interface of membrane binding domains
Were acetylation to play a widespread role in controlling protein-membrane interaction and
function, one might expect acetylation sites to be concentrated in membrane-binding regions of
membrane-binding proteins. To test this notion, we developed a bioinformatics approach that
compares the location of PTMs to high-resolution structural information of membrane-binding
proteins with known membrane-binding surfaces. We chose four structurally well-characterized
families of membrane-binding domains where acetylation data were available for many of the
family members at Phosphosite.org (4), bin/amphiphysin/rvs (BAR), phox-homology (PX), C2,
and Eps15-homology (EHD). We defined membrane-binding interfaces using structural and/or
sequence alignments and determined the likelihood of finding acetylated lysine residues within or
outside the defined membrane-binding region (schematically illustrated in Figure A.S1). As
summarized in table A.1, acetylation was predominantly localized to the membrane-binding
regions of all domain families analyzed. For BAR domains, acetylation was preferentially
localized to the membrane-binding region by an order of magnitude. We see acetylation primarily
confined on the concave surface of most BAR proteins. Unlike most BAR proteins (5), I-BAR
protein, IRSp53, binds membranes on its convex surface, which is where acetylation is localized.
Twelve of thirteen BAR (Figure A.S2) and eleven of thirteen PX domains analyzed (Figure A.S3)
localized acetylation exclusively to their membrane-binding regions. Likewise, eleven of thirteen
C2A domains and all nine C2B domains analyzed, confined acetylation only to membrane-binding
regions (Figure A.S4). A similar pattern of acetylation emerged for EHD family members (Figure
A.S5). Across all of the domains analyzed, we found 20.40% compared to 3.05% of lysines
142

acetylated within membrane-binding regions verses outside. This preferential localization of
acetylation to membrane-binding regions across multiple families of domains is consistent with
acetylation playing an important, general role in controlling membrane protein function.
Table A.1

Table A.1 – Prevalence of Lysine Acetylation by domain in membrane and non-membrane binding
regions of BAR, PX, C2 and EHD membrane binding domains and their subfamilies. K
AC
, number
of acetylated lysine residues; K
TOT
, number of lysine residues; Mem, membrane binding region;
Non-mem, non-membrane binding region.
143

FigureA.S1

Figure A.S1 – Algorithm of the bioinformatics approach used to determine the differential
distribution of acetylation within membrane binding domains. MBR, Membrane binding region;
K
AC
, Acetylated lysine; K
TOT
, total number of lysines within a defined region.

144

Figure A.S2

Figure A.S2 – Acetylation within BAR domains is localized to membrane binding regions. A)
BIN2 (4AVM.pdb) structure displayed to illustrate membrane binding residues (blue α-carbons)
and non-membrane binding residues (gray α-carbons). A red line approximates the level of lipid
headgroups. B) BAR domain acetylation prevalence listed by associated BAR domain containing
gene. Percentages represent the number of acetylated lysine residues out of the total number of
lysines in their respective regions. (.pdb) file names indicate the PDB crystal structure used in our
analysis.
145

Figure A.S3

146

Figure A.S3 – Acetylation within PX domains is localized to membrane binding regions. A)
p40
phox
structure (1H6H.pdb) with membrane binding regions (blue) and non-membrane binding
regions (gray), as determined by previous studies (6–9). The red line approximates the level of
lipid headgroups. B) Sequences of PX domains aligned against p40phox (residues 19-140). The
blue bar denotes the membrane binding region. Acetylation sites are marked in green. C)
Acetylation prevalence among PX domain containing genes. Percentages represent the number of
acetylated lysines out of the total lysines in their respective regions. (.pdb) file names indicate the
PDB crystal structure used.
147

Figure A.S4

148

Figure A.S4 – Acetylation within C2 domains is localized to membrane binding regions. A & B)
Synaptotagmin1 C2A and C2B domain structures (2R83.pdb) with Ca
2+
and membrane binding
regions highlighted blue. C & D) Sequence alignments of C2A (157-245) and C2B (287-378)
domains are shown with blue bars denoting membrane binding residues and green boxes around
acetylation sites. E) Acetylation prevalence in membrane and non-membrane binding regions are
shown by gene. (.pdb) file names indicate the PDB crystal structure used.
149

Figure A.S5

Figure A.S5 – Eps15-homology (EHD) 1-4 domains localize acetylation to their membrane
binding regions. A) The crystal structure of EHD2 (4CID.pdb) with membrane binding regions
labeled in blue. The red line represents the level of the lipid headgroups. B) Sequence alignments
150

of EHDs 1, 3, and 4 against EHD2 (300-417). Membrane binding regions are demarcated by a blue
bar and acetylated lysines are denoted by a green box. C) Prevalence of acetylation in membrane
and non-membrane binding regions for EHD1-4.
A.2.2 Mimicking lysine acetylation in Amphiphysin inhibits membrane remodeling
by reducing membrane binding affinity and dissociates amphiphysin from tubular
networks in cells
To test whether acetylation modulates membrane interactions we investigated three well-
characterized membrane-binding protein candidates from the BAR, C2 and EHD families,
amphiphysin (10), synaptotagmin1 (11) and EHD2 (12), respectively. To recapitulate acetylation,
recombinant constructs were generated with key lysines (K) mutated to glutamines (Q), a common
method for mimicking acetylation that, like acetylation, results in a loss the positive lysine charge
in lieu of an amide bond (13).
The N-terminus of amphiphysin’s BAR domain is important for inducing membrane curvature and
has acetylation sites at positions 5 and 15 (4)
,
(10)
,
(14)
,
(15). Using transmission electron
microscopy (TEM), we first studied the ability of acetylation mimicking mutations at K5Q-K15Q
in amphiphysin (Amph-5Q15Q) to remodel vesicles. Incubation of wild type amphiphysin
(Amph-WT) with a lipid composition optimized for membrane tubulation (POPG:POPE) (10)
results in near complete tubulation of liposomes, while incubation with Amph-5Q15Q results in
mostly vesicles and beaded structures (Figure A.1 a,b). To test whether the effect of acetylation
was due to the change in charge or side chain structure, we replaced positions 5 and 15 with
arginine residues (Amph-5R15R) and found that tubulation was similar between the wild type and
arginine mutant (Figure A.1 c). When incubated with a lipid composition designed to better mimic
cellular membranes (POPS:POPC), only Amph-WT and Amph-5R15R generated tubules (Figure
151

A.1 d-f). It is noteworthy that while little to no remodeling is observable in POPS:POPC lipid
treated with Amph-5Q15Q, POPG:POPE lipid similarly treated is still remodeled, but the
tubulation phenotype is lost in favor of vesicles and beaded structures.
Next, we tested whether the inability to tubulate correlated with a change in membrane binding
affinity. We employed an electron paramagnetic resonance (EPR)-based method that measures
amplitude changes upon membrane binding during lipid titration (16). A right shift in the titration
curve of Amph-5Q15Q compared with Amph-WT was evident, demonstrating that acetylation
attenuates protein-membrane binding affinity (Figure A.1 j). These data are consistent with the
hypothesis that acetylation in the membrane-binding regions of amphiphysin alters its ability to
remodel lipids via modulation of membrane affinity.
We next sought to determine whether acetylation-dependent loss of membrane-binding affinity
and elicited changes in cellular phenotypes. Toward this end, we developed C-terminal green-
fluorescent protein labeled constructs of amphiphysin wild type (wt-Amph-GFP), K5Q K15Q
double mutant (5Q15Q-Amph-GFP), and K15Q single mutant (15Q-Amph-GFP). Consistent with
prior findings (15), we found wt-Amph-GFP formed tubular networks in COS-7 cells (Figure A.1
g). Conversely, neither 5Q15Q-Amph-GFP nor 15Q-Amph-GFP were associated with a tubular
network and were instead primarily diffuse throughout the cytoplasm (Figure A.1 h, i). This
altered localization of the 5Q15Q-Amph-GFP and 15Q-Amph-GFP mutants, is consistent with the
loss of affinity and inability to tubulate membranes observed using liposomes.
152

Figure A.1

Figure A.1 – Acetylation modulates amphiphysin-membrane interaction properties. A-C)
POPG:POPE lipids incubated with Amph-WT (a), Amph-5Q15Q (b), or Amph-5R15R (c)
153

visualized by TEM. D-F) POPS:POPC lipids incubated with Amph-WT (d), Amph-5Q15Q (e), or
Amph-5R15R (f). G-I) Fluorescent confocal microscopy of wt-Amph-GFP (g) 5Q15Q-Amph-
GFP (h) 15Q-Amph-GFP expressed for 24 hours in COS-7 and imaged live. J) EPR spectral
amplitudes as a function of lipid concentration from Amph-WT (solid line) and Amph-5Q15Q
(dashed line) spin-labeled at position 20. Error bars represent 1 standard deviation from at least 3
independent measurements. Scale bars: black = 500 nm, white = 10 µm.
A.2.3 Mimicking lysine acetylation in Eps Homology Domain 2 prevents lipid
remodeling, reduces membrane binding affinity and leads to dissociation from the
plasma membrane
To determine whether acetylation similarly modulates EHD2 interactions with membranes, we
again employed TEM to visualize EHD2 membrane interactions. EHDs bind and tubulate
membranes, and in the case of EHD2 this enhances ATPase activity
~
8 fold (17). To mimic
acetylation, we mutated the K324 (4)
,
(18) acetylation site to glutamine (EHD2-324Q). K324
penetrates deeply into the bilayer when EHD2 binds membranes (12). As expected wild type
EHD2 (EHD2-WT) displayed a strong tubulating phenotype (Figure A.2a). By comparison, the
ability of EHD2-324Q to tubulate lipids was dramatically reduced (Figure A.2b). In addition, a
significant decrease in membrane binding affinity was observed for EHD2-324Q compared to
EHD2-WT when analyzed for lipid binding in both EPR and vesicle co-sedimentation assays
(Figure A.2 c, g). A corresponding decrease in ATPase activity was observed for EHD2-324Q
(Figure A.2h) consistent with dissociation from the membrane. Thus, mimicking acetylation
reduces EHD2’s membrane affinity, membrane remodeling capability and catalytic activity.
In HeLa cells N-terminally GFP-tagged EHD2 (GFP-EHD2-wt) distributes to the cellular
membrane and the cytoplasmic region immediately adjacent (17) (Figure A.2d). GFP-EHD2-324Q
154

or GFP-EHD2-324Q328Q (both acetylation sites) (4,18) produced diffuse cytoplasmic
distributions (Figure A.2 e, f) further illustrating the altered membrane binding behavior of EHD2
with acetylation mimetics in its membrane-binding region.
Figure A.2

Figure A.2 – Acetylation modulates EHD2-membrane interactions. A, B) TEM of EHD2-WT (a)
or EHD2-324Q (b) incubated with Folch lipids. C) EHD2-WT and EHD2-324Q centrifuged with
155

or without Folch-LUVs. D-F) HeLa cells expressing wt-EHD2-GFP (d), 324Q-EHD2-GFP (e), or
324Q328Q-EHD2-GFP (f). G) EPR spectral amplitudes plotted as a function of lipid concentration
from EHD2-WT (solid line) and EHD2-324Q (dashed line) spin-labeled at position 321. H) ATP
consumption by EHD2-WT and EHD2-324Q in the presence or absence of lipid. Error bars
represent ±1 SD from ≥ 3 experiments. Scale bars: black = 500 nm, white = 10 µm. SN =
Supernatant. P = Pellet.
A.2.4 Mimicking lysine acetylation in the C2A domain of Synaptotagmin 1 reduces
membrane binding affinity
Using a similar approach, we examined whether the acetylation mimicking K237Q mutant in the
membrane-binding region of the C2 domain-containing protein synaptotagmin1 (Syt1-237Q)
could alter membrane interaction. Again, we found the lipid titration curve for Syt1-237Q right-
shifted compared to wild type controls, indicating reduced membrane binding affinity (Figure
A.S6).
Figure A.S6

156

Figure A.S6 – Acetylation reduces synaptotagmin1 membrane binding affinity. EPR spectral
amplitudes are plotted as a function of lipid concentration for Syt1-WT (solid line) and Syt1-237Q
(dashed line) spin-labeled at position 227. Error bars represent ±1 SD from ≥ 3 experiments.
A.2.5 Mimicking lysine acetylation of Amphiphysin disrupts the T-tubule network
and causes a flight defect in D. melanogaster
Next we sought to verify our findings using an in vivo system. Amphiphysin stabilizes the T-tubule
network in Drosophila melanogaster muscle tissue (19). To test whether mimicking acetylation of
amphiphysin alters its function in vivo, Amph-WT and Amph-5Q15Q constructs were introduced
into an amphiphysin-null (Amph26) background and slices of muscle tissue as well as flight
behavior were analyzed. Normal flies exhibit a characteristic pattern of disc large (DLG) staining,
a determinant of proper T-tubule formation ((19)), severely disrupted in Amph26 flies (Figure A.3
a). Expression of Amph-WT rescued T-tubule formation while Amph-5Q15Q yielded severely
disrupted T-tubules similar to Amph26,

despite comparable expression levels (Figure A.3 a, b).
Amph26

flies display a flight defect consistent with T-tubule disruption. Whereas this defect is
ameliorated by Amph-WT expression, Amph-5Q15Q expressing flies exhibited a strong flight
deficit (Figures A.3 c). The T-tubule degeneration and the flight defect phenotype of the
acetylation mimicking mutant were in excellent agreement with the biochemical and cell data,
which revealed a strongly reduced ability to stabilize tubular structures.
157

Figure A.3

Figure A.3 – Amph-5Q15Q flies display defective T-tubule organization and a flight deficit. A)
Representative images of adult IFM stained for actin (phalloidin, red) and DLG (green), with
Amph-WT

or Amph-5Q15Q expression driven by 24B-GAL4 over an amphiphysin-null (Amph-
26) genetic background. B) Western blot demonstrating levels of transgene expression detected
158

using an antibody against the SH-3 domain of Amphiphysin. C) Percentage of flies that are flight
defective. n is reported at the base of each condition where, n = number of flies tested. Error bars
represent ±1 SD from ≥ 3 experiments. Scale bar = 2 μm.
A.3 DISCUSSION
Lysine acetylation is an important and common (20) regulatory mechanism. Here we show several
new lines of evidence to suggest that acetylation plays a central role in controlling membrane
protein function. The finding that acetylation is strongly enhanced in membrane-binding regions
implies functional specificity for acetylation in the control of protein-membrane interactions.
Consistent with acetylation blocking key charge interactions between protein and membrane,
mimicking acetylation in amphiphysin, Syt1 and EHD2 reduced membrane binding affinity, in
essence removing them from the membrane (Figure A.S7). Interestingly, this had differential
effects on membrane remodeling, thus, depending on circumstance, acetylation can block or alter
protein-membrane interactions. However, in cells, mimicking even a single acetylation event
sufficiently reduced affinity with membranes enough to induce cytoplasmic dispersion. As lysine
acetylation is a reversible modification, its effect on protein-membrane interactions could be easily
toggled on or off at specific times and places during membrane remodeling processes within the
cell. Here we tested proteins with known membrane-remodeling functions and observed
biophysical, cellular, tissue and behavioral level consequences, yet our data imply broad
applicability to controlling a wide array of membrane protein functions. For instance, acetylation
of lysines K237 and K302 in the C2A and C2B Ca
2+
coordination site of Syt1 are uniquely placed
to regulate exocytosis (21). Acetylation of p40
phox
at K92 would block critical interactions with
the 1-phosphate of phosphatidylinositol 3-phosphate, important in recruiting the NADPH oxidase
complex to the membrane for oxidative burst of neutrophils (6,22). Furthermore, PTMs such as
159

phosphorylation can change a protein’s membrane-binding qualities, for instance, by altering depth
of insertion into the membrane (16). Intriguingly, phosphorylation sites can often be found in the
vicinity of acetylation sites in membrane-binding regions. For example, simultaneous acetylation
of endophilin A1 at K76, adjacent to the S75 phosphorylation site that functions as a switch
between membrane binding modes, could have significant effects on the endocytosis process that
endophilin controls (16,23). Thus, acetylation in membrane binding regions could yield fascinating
consequences to both the number of functional modes a membrane-binding protein has and the
interplay between cellular systems controlling PTM events. Future studies will be required to
uncover the cellular machinery that brings about acetylation and deacetylation of lysines in
membrane binding regions.
160

Figure A.S7

Figure A.S7 – Cartoon model of lysine acetylation of a membrane binding residue inducing a loss
of membrane affinity between a membrane binding protein and its target membrane. Charge-
charge interactions between key lysine residues on membrane binding proteins with negatively
161

charged lipid headgroups are crucial for protein-membrane interactions. Acetylation of lysine
residues will mask the charged amine group, blocking charge-charge interactions between protein
and membrane, reducing membrane binding affinity. As a result acetylated proteins can leave the
membrane. Purple, membrane binding protein; yellow, membrane-binding lysine; blue sphere,
neutral lipid headgroup; red sphere, negatively charged lipid headgroup.
162

A.4 REFERENCES
1. Gräff J, Tsai L-H. Histone acetylation: molecular mnemonics on the chromatin. Nat Rev
Neurosci. 2013 Feb;14(2):97–111.
2. Choudhary C, Weinert BT, Nishida Y, Verdin E, Mann M. The growing landscape of lysine
acetylation links metabolism and cell signalling. Nat Rev Mol Cell Biol. 2014
Aug;15(8):536–50.
3. McMahon HT, Gallop JL. Membrane curvature and mechanisms of dynamic cell membrane
remodelling. Nature. 2005 Dec 1;438(7068):590–6.
4. Hornbeck PV, Kornhauser JM, Tkachev S, Zhang B, Skrzypek E, Murray B, et al.
PhosphoSitePlus: a comprehensive resource for investigating the structure and function of
experimentally determined post-translational modifications in man and mouse. Nucleic
Acids Res. 2012 Jan;40(Database issue):D261-270.
5. Saarikangas J, Zhao H, Pykäläinen A, Laurinmäki P, Mattila PK, Kinnunen PKJ, et al.
Molecular mechanisms of membrane deformation by I-BAR domain proteins. Curr Biol
CB. 2009 Jan 27;19(2):95–107.
6. Bravo J, Karathanassis D, Pacold CM, Pacold ME, Ellson CD, Anderson KE, et al. The
crystal structure of the PX domain from p40(phox) bound to phosphatidylinositol 3-
phosphate. Mol Cell. 2001 Oct;8(4):829–39.
7. Lee SA, Kovacs J, Stahelin RV, Cheever ML, Overduin M, Setty TG, et al. Molecular
mechanism of membrane docking by the Vam7p PX domain. J Biol Chem. 2006 Dec
1;281(48):37091–101.
163

8. Lomize AL, Pogozheva ID, Lomize MA, Mosberg HI. The role of hydrophobic interactions
in positioning of peripheral proteins in membranes. BMC Struct Biol. 2007 Jun 29;7(1):44.
9. Stahelin RV, Burian A, Bruzik KS, Murray D, Cho W. Membrane Binding Mechanisms of
the PX Domains of NADPH Oxidase p40 phox and p47 phox. J Biol Chem. 2003 Apr
18;278(16):14469–79.
10. Isas JM, Ambroso MR, Hegde PB, Langen J, Langen R. Tubulation by amphiphysin
requires concentration-dependent switching from wedging to scaffolding. Struct Lond Engl
1993. 2015 May 5;23(5):873–81.
11. Honigmann A, van den Bogaart G, Iraheta E, Risselada HJ, Milovanovic D, Mueller V, et
al. Phosphatidylinositol 4,5-bisphosphate clusters act as molecular beacons for vesicle
recruitment. Nat Struct Mol Biol. 2013 Jun;20(6):679–86.
12. Shah C, Hegde BG, Morén B, Behrmann E, Mielke T, Moenke G, et al. Structural insights
into membrane interaction and caveolar targeting of dynamin-like EHD2. Struct Lond Engl
1993. 2014 Mar 4;22(3):409–20.
13. Zhou B-R, Feng H, Ghirlando R, Kato H, Gruschus J, Bai Y. Histone H4 K16Q mutation,
an acetylation mimic, causes structural disorder of its N-terminal basic patch in the
nucleosome. J Mol Biol. 2012 Aug 3;421(1):30–7.
14. Zhao S, Xu W, Jiang W, Yu W, Lin Y, Zhang T, et al. Regulation of cellular metabolism
by protein lysine acetylation. Science. 2010 Feb 19;327(5968):1000–4.
164

15. Peter BJ, Kent HM, Mills IG, Vallis Y, Butler PJG, Evans PR, et al. BAR domains as sensors
of membrane curvature: the amphiphysin BAR structure. Science. 2004 Jan
23;303(5657):495–9.
16. Ambroso MR, Hegde BG, Langen R. Endophilin A1 induces different membrane shapes
using a conformational switch that is regulated by phosphorylation. Proc Natl Acad Sci U
S A. 2014 May 13;111(19):6982–7.
17. Daumke O, Lundmark R, Vallis Y, Martens S, Butler PJG, McMahon HT. Architectural
and mechanistic insights into an EHD ATPase involved in membrane remodelling. Nature.
2007 Oct 18;449(7164):923–7.
18. Choudhary C, Kumar C, Gnad F, Nielsen ML, Rehman M, Walther TC, et al. Lysine
acetylation targets protein complexes and co-regulates major cellular functions. Science.
2009 Aug 14;325(5942):834–40.
19. Razzaq A, Robinson IM, McMahon HT, Skepper JN, Su Y, Zelhof AC, et al. Amphiphysin
is necessary for organization of the excitation–contraction coupling machinery of muscles,
but not for synaptic vesicle endocytosis in Drosophila. Genes Dev. 2001 Nov
15;15(22):2967–79.
20. Prabakaran S, Lippens G, Steen H, Gunawardena J. Post-translational modification: nature’s
escape from genetic imprisonment and the basis for dynamic information encoding. Wiley
Interdiscip Rev Syst Biol Med. 2012 Dec;4(6):565–83.
21. Corbalan-Garcia S, Gómez-Fernández JC. Signaling through C2 domains: more than one
lipid target. Biochim Biophys Acta. 2014 Jun;1838(6):1536–47.
165

22. Wientjes FB, Hsuan JJ, Totty NF, Segal AW. p40phox, a third cytosolic component of the
activation complex of the NADPH oxidase to contain src homology 3 domains. Biochem J.
1993 Dec 15;296 ( Pt 3):557–61.
23. Boucrot E, Ferreira APA, Almeida-Souza L, Debard S, Vallis Y, Howard G, et al.
Endophilin marks and controls a clathrin-independent endocytic pathway. Nature. 2015 Jan
22;517(7535):460–5.
24. Hornbeck PV, Kornhauser JM, Tkachev S, Zhang B, Skrzypek E, Murray B, et al.
PhosphoSitePlus: a comprehensive resource for investigating the structure and function of
experimentally determined post-translational modifications in man and mouse. Nucleic
Acids Res. 2012 Jan;40(Database issue):D261-270.
25. Sievers F, Wilm A, Dineen D, Gibson TJ, Karplus K, Li W, et al. Fast, scalable generation
of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol.
2011;7:539.
26. Robert X, Gouet P. Deciphering key features in protein structures with the new ENDscript
server. Nucleic Acids Res. 2014 Jul;42(Web Server issue):W320-324.
27. Fuson KL, Montes M, Robert JJ, Sutton RB. Structure of human synaptotagmin 1 C2AB in
the absence of Ca2+ reveals a novel domain association. Biochemistry (Mosc). 2007 Nov
13;46(45):13041–8.
28. Coppola T, Magnin-Lüthi S, Perret-Menoud V, Gattesco S, Schiavo G, Regazzi R. Direct
Interaction of the Rab3 Effector RIM with Ca2+Channels, SNAP-25, and Synaptotagmin. J
Biol Chem. 2001 Aug 31;276(35):32756–62.
166

29. Isas JM, Ambroso MR, Hegde PB, Langen J, Langen R. Tubulation by amphiphysin
requires a concentration-dependent switching from wedging to scaffolding. Structure.
2015;in press.
30. Peter BJ, Kent HM, Mills IG, Vallis Y, Butler PJG, Evans PR, et al. BAR domains as sensors
of membrane curvature: the amphiphysin BAR structure. Science. 2004 Jan
23;303(5657):495–9.
31. Shah C, Hegde BG, Morén B, Behrmann E, Mielke T, Moenke G, et al. Structural Insights
into Membrane Interaction and Caveolar Targeting of Dynamin-like EHD2. Struct Lond
Engl 1993. 2014 Mar 4;22(3):409–20.
32. Frazier AA, Roller CR, Havelka JJ, Hinderliter A, Cafiso DS. Membrane-Bound
Orientation and Position of the Synaptotagmin I C2A Domain by Site-Directed Spin
Labeling†. Biochemistry (Mosc). 2003 Jan 1;42(1):96–105.
33. Banerjee S, Lee J, Venkatesh K, Wu C-F, Hasan G. Loss of flight and associated neuronal
rhythmicity in inositol 1,4,5-trisphosphate receptor mutants of Drosophila. J Neurosci Off J
Soc Neurosci. 2004 Sep 8;24(36):7869–78.

A.5 AUTHOR CONTRIBUTION STATEMENT
Author contributions: M.R.A., A.O., O.D., K.C., I.H., and R.L. contributed to the conception and
design of this study. M.R.A., A.O., J.L., A.M, D.M., K.T., K.C., O.D., I.H., performed
experiments. A.O., M.R.A., J.L., I.H., and R.L. wrote the manuscript. M.R.A., A.O., J.L., A.M,
K.T., K.C., O.D., I.H., and R.L. edited the manuscript.
167

A.6 METHODS
Reagents
1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-RAC-(1-glycerol)] (POPG), 1-palmitoyl-2-oleoyl-
sn-glycero-3-phosphoethanolamine (POPE), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-
serine (POPS) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) were obtained
from Avanti Polar Lipids (Alabaster, AL). Bovine total brain lipids, Folch fraction I was obtained
from Sigma-Aldrich (St. Louis, MO). 1-oxyl-2,2,5,5 tetramethyl-Δ3-pyrroline-3-
methylmethanethiosulfonate (MTSL) was purchased from Toronto Research Chemicals (Toronto,
Ontario, Canada).
Experimental Procedures
Bioinformatic analysis of protein acetylation. The bioinformatics approach was composed of
distinct steps schematically represented in (Extended Data 1). Membrane binding domain families
were first chosen. In selecting domain, families we focused on characteristically static domains
with conserved structural or sequence homology (BAR, PX, C2 and EHD) in order to allow us to
generalize membrane binding properties across family members. In addition, the availability of
detailed 3D structural data and PTM data was considered. Once the domain families were chosen,
domains lacking acetylation data were excluded. If crystal structures for domains were not
available, the domain was sorted for sequence alignment analysis. If a crystal structure was
available, then the domain was sorted for structural alignment instead. A template membrane
binding domain was chosen based on experimental data and the cohort of domains, including the
template domain, was sorted into clusters based on structural homology. Domains with sufficient
structural homology were then aligned against the template domain and acetylation sites were
168

determined to be within or outside the membrane binding region of the domain. These data were
normalized by the total number of lysines within or outside the membrane binding region to
differentiate the frequency of lysine acetylation at the membrane binding interface from the
frequency of acetylation in non-membrane binding regions. The data are presented by domain in
Table 1 and by protein in extended data figures 2 – 5 (in some cases single proteins had multiple
membrane binding domains).
Acetylation data was acquired from www.phosphosite.org (24) in December of 2014. Structure
files (www.rcsb.org) for membrane binding domain families were downloaded in bulk. Scripts
were generated to correlate the modification data with .pdb files and sort these files into homology
clusters based on a structural homology algorithm we generated. Family members with less than
50% sequence identity were excluded from the analysis. 3D structures were aligned in PyMOL
(PyMOL Molecular Graphic System, Version 1.7.4 Schrödinger, LLC) where identification of the
acetylation sites present in membrane or non-membrane binding regions was made. False positives
were eliminated during this step of the analysis due to their lack of structural homology. In
comparing structural overlays between template domains and crystal structures from homologous
family members, a 2 Å maximum distance from an overlaid template residue defined as membrane
binding was used as cutoff criteria for inclusion of an acetylated lysine within a membrane binding
region.
Sequence alignments were performed using ClustlOmega (25) and ESPript-http://espript.ibcp.fr
(26). Acetylation sites from Phosphosite.org (24) were manually annotated. Alignments were
made against a family member with a membrane binding region well-defined by empirical data
available in the literature with a determined 3D structure, p40
phox
, Syt1 and EHD2 for PX, C2 and
EH domains, respectively (6,12,27).
169

The tallies of acetylated lysines in membrane or non-membrane binding regions determined by
structural or sequence homology were normalized to the total number of lysines in each membrane
or non-membrane binding region.
Protein expression, DNA constructs and mutagenesis. Plasmids and cDNA encoding rat
synaptotagmin1 (a.a. 80-421), and Drosophila amphiphysin (full-length and BAR domain; a.a. 1-
244) were the generous gifts from Drs. Greg Schiavo and Harvey McMahon respectively.
Constructs made for recombinant expression, including mouse EHD2 (a.a. 1-404) and all mutants
were purified as previously described (28–31). Briefly, proteins were expressed in E. coli BL21
(DE3) (New England Biolabs). EHD2 and amphiphysin were purified using nickel-nitrilo-triacetic
acid–agarose, followed by gel filtration with a Superdex 200 column, and in some cases remaining
impurities were removed using mono S cation exchange chromatography with a low salt buffer A
(20 mM hepes pH 7.4, 1 mM dithiothreitol (DTT)) and elution buffer B (20 mM hepes pH 7.4, 2
M NaCl and 1 mM DTT). Synaptotagmin1 was purified by immobilizing the protein on
glutathione-agarose followed by extensive washing. Synaptotagmin1was eluted off the beads by
thrombin cleaving of the GST tag from the protein. Protease and other impurities were removed
using a mono Q column (GE). Protein concentrations were determined by UV-absorbance at 280
nm. The purified samples were flash frozen and stored at -80° C.
Lysine to glutamine or arginine mutations were made following Quikchange (Agilent) site-
directed mutagenesis manufacturer protocols. In order to allow specific spin-labeling of sites on
amphiphysin, EHD2 and synaptotagmin1 for EPR experiments, native cysteine residues were
mutated to alanines as before (29,31,32) to create cys-less versions of each protein and site-specific
mutation of cysteines was performed in locations known to embed in membranes. In the case of
170

amphiphysin, position 20; for EHD2, position 321; and for synaptotagmin1, position 227 were
chosen. Spin label was incubated in a 5- to 10-fold molar excess of protein immediately following
the removal of DTT using size exclusion chromatography (PD-10 column (GE)), and left to react
at 4° C overnight. Excess spin label was removed using PD-10 columns.
Vesicle preparation, tubulation assays and electron microscopy. The initial preparation of
vesicles was the same for all lipid compositions used in this study. Lipid stocks were suspended in
chloroform and mixed to the desired molar or weight proportions in organic solvent, dried under
a stream of N2 gas, and dried overnight in a desiccator. For tubulation TEM studies with
amphiphysin and EHD2, multilamellar vesicles (MLVs) of 2:1 wt/wt POPG and POPE were
resuspended in buffer A (20 mM Hepes, pH 7.4) to 4 mg/mL and used immediately. Amphiphysin
was also tested for its interaction with MLVs of 2:1 wt/wt POPS and POPC, which were
resuspended in buffer A to 4 mg/mL. Protein and lipid were mixed at a 1:70 and 1:2 protein:lipid
molar ratios for amphiphysin and EHD2, respectively. Carbon-coated formvar films mounted on
copper grids (Electron Microscopy Services, Hatfield) were suspended on small aliquots of
samples for 10 minutes. Excess liquid was removed using filter paper and grids were subsequently
subjected to a two-minute incubation on a droplet of 1% uranyl acetate was used to stain the
sample-coated grids. A JEOL 1400 transmission electron microscope was used for specimen
observation at 100 kV.
Liposome co-sedimentation assay. Sedimentation assays were performed as before (31). Briefly,
10 μM EHD2 in 20 mM HEPES (pH 7.5), 300 mM NaCl, and 0.5 mM MgCl2 at RT was incubated
with Folch-LUVs at 1 g/l for 10 min and centrifuged at 200,000 × g for 10 min in a TLA 100 rotor.
SDS-PAGE was used to analyze the pellet and supernatant.
171

ATPase assays. ATPase assays were performed as before (17). Briefly, assays were conducted in
20 mM HEPES (pH 7.5), 135 mM NaCl, 15 mM KCl, 1 mM MgCl2 at 30 °C with 10 μM enzyme
(EHD2-WT or EHD2-324Q) and 100 μM ATP as substrate. In some cases, 1 mg/ml sonicated
Folch fraction I (Sigma-Aldrich) liposomes were added. Addition of the enzyme to the reaction
mixtures initiated enzymatic activity at which point nucleotide hydrolysis was determined using
HPLC measurement. A linear fit applied to data up to 40% nucleotide hydrolysis was used to
determine initial rates of hydrolysis.
Acquisition and analysis of EPR data in lipid binding assays. Continuous wave (CW) EPR
spectra were recorded for samples placed into Quartz capillaries (VitroComInc., New Jersey) using
a Bruker EMX spectrophotometer fitted with an ER4119HS resonator. For lipid titration
experiments of amphiphysin, MLVs composed of 2:1 wt/wt POPG:POPE were used. For EHD2,
a previously optimized lipid system of bovine total brain lipids, Folch I (Sigma) (31) were used.
These lipids were suspended in 20 mM Hepes, pH 7.4, 100 mM NaCl, and made into small
unilamellar vesicles (SUVs) using a tip sonicator. Synaptotagmin1 was previously studied by EPR
on vesicles composed of a 3:1 molar ratio of POPC:POPS, and extruded through polycarbonate
filters with a 0.1 μm pore diameter (32). CW EPR spectral amplitudes were recorded for samples
of spin labeled protein in 20 mM Hepes, pH 7.4, 100 mM NaCl buffer, (as well as 1 mM Ca
2+
in
the case of synaptotagmin1). For all experiments, protein concentration was 10 μM and the amount
of lipids added was varied. The values were then normalized relative to the protein’s CW EPR
spectral amplitude difference between the protein alone in solution and at saturating conditions.
Unlike the spin labeled constructs of amphiphysin and EHD2, synaptotagmin1 spin labeled at
position 227 did not undergo major changes in signal amplitude, similar to what was observed for
this region of synaptotagmin1 previously (32). However, increasing concentrations of lipids
172

significantly increased an immobilized component in the low field transition line. Therefore, the
ratio of the amplitude of this peak and that of the central line width was used to plot the effect of
increasing lipid concentrations and normalized to its highest (saturating conditions) and lowest (in
solution) values obtained.
Cell culture, Transfection and Confocal Microscopy
For expression of mutant and wild type amphiphysin in eukaryotic systems, the N-BAR domain
was cloned into a pEGFP-N1 vector using 5’ NheI and 3’ XhoI cut sites. For expression of mutant
and wild type EHD2 in eukaryotic systems an N-terminally GFP-tagged construct was used, as
before (31).
HeLa and COS-7 cell lines were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM)
supplemented with 100 U/mL penicillin G, 100 μg/mL streptomycin, 4.5 g/L glucose, sodium
pyruvate (Cellgro, Manassas, VA) and 10% heat inactivated fetal bovine serum (Invitrogen) at 37
°C in humidified at with 5% CO2.
Following three washes with phosphate buffered saline solution, HeLa or COS-7 cells lines were
trypsin digested and the cell suspensions were centrifuged at 1,000 x g for 5 minutes. Cells were
recovered in fresh media and plated on custom, #1 thickness, glass-bottomed coverslips and
allowed to recover for 24 hours before transfection.
Expression of cDNA constructs was induced using Lipofectamine 2000 (Invitrogen) and 1.2 μg of
cDNA plasmid according to manufacturer protocol. Cells were imaged at 24 hours following
transfection with an Olympus IX-83 confocal microscope using an UPLFN 100x oil immersion
objective (NA: 1.30). eGFP fluorescence was excited using a 488 nm laser and light was collected
through the objective.
173

Images were acquired for analysis, which was performed using ImageJ software from the NIH
(version 1.48).
Drosophila Melanogaster Studies
For expression of amphiphysin in AMPH26 flies, full length amphiphysin with or without
acetylation mutations was sub-cloned into the pINDY6 vector using 5’ XhoI and 3’ SpeI restriction
sites. and used to generate flies that carried WT or 5Q15Q-Amphiphysin. Subsequent crosses with
AMPH26 flies allowed expression of the WT or acetylation mimic in the genetic background of
the knockout.
Western Blotting
To detect Amph levels in flies, protein extracts were obtained by homogenizing flies in RIPA lysis
buffer (50 mM Tris-HCl, pH7.5, 1% NP-40, 0.5% NaDoc, 150 mM NaCl, 0.1% SDS, 2 mM
EDTA, 50 mM NaF, 1 mM Na3VO4, 250 nM cycloporin A, protease inhibitor cocktail (Roche)
and phosphatase inhibitor cocktail 1 (Sigma) using mortar and pestle. 15 μg protein homogenate
was separated by SDS-PAGE and transferred to nitrocellulose membrane. Primary antibodies were
diluted in blocking solution as following: rabbit anti-Amph SH3 domain (from Dr. Harvey
McMahon) 1:1500; anti-tubulin 1:500 (7E10, DSHB).
Immunocytochemistry
Indirect flight muscles, IFM, were dissected in PBS and fixed in 4% paraformaldehyde for 25 min.
Fixed samples were washed with 0.1% triton X-100 in PBS (PBST) then blocked with 5% normal
goat serum (NGS) in PBST. Mouse anti-DLG 1:100 (4F3, DSHB) was diluted in blocking buffer.
Alexa-conjugated secondary antibody was used at 1:250 (Invitrogen). 100 M working stock of
fluorescent actin-stain phalloidin (Cytoskeleton) was used to stain actin filaments. Images were
174

captured using Zeiss LSM5 confocal microscope using a 63X 1.6NA oil immersion objective with
a 2x zoom. When comparing intensity across genotypes, the exposure time was kept constant for
all genotypes per experiment.
Flight tests
An adapted “cylinder drop” flight assay was performed similar to that described by Banerjee et al.
(2004) (33). A transparent flight chamber with 8-cm inner diameter was made from a plastic
transparent sheet. A wider tube that could hold the narrower “drop tube” was attached to the funnel
and was placed on top of the flight chamber. 2-4 day old flies were transferred to narrow plastic
“drop tube” and 10 to 20 flies per vial were tested per experiment. Assembled flight chamber was
placed on ice to count the number of flies that dropped directly down the chamber, which were
considered as defective in flight. The percentage of flight defective flies was calculated by
counting the number of flies that dropped to the bottom, divided by the total number of flies tested,
and then multiplied by 100. At least 5 independent assays were performed per genotype.

175

CONCLUDING REMARKS
Protein misfolding diseases represent a unique class of human disease. Despite being among the
most common and debilitating chronic conditions seen in the modern era, diseases like type 2
diabetes mellitus, Alzheimer disease and Parkinson disease have yet to see therapeutics developed
that can stop disease progression. A wealth of studies over the past 50 years have helped to place
the importance of protein misfolding at the center of many protein-misfolding disease pathogenic
mechanisms. The development and refinement of protein structural analysis tools has helped to
make this possible, including X-ray crystallography, electron paramagnetic resonance
spectroscopy, liquid and solid state nuclear magnetic resonance spectroscopy, transmission and
cryo-electron microscopy, Fourier transform infrared spectroscopy, and circular dichroism.
Furthermore, the formalization and application of chemical reaction kinetics methods for use in
amyloid aggregation studies has allowed for quantitative investigations into the protein misfolding
process to be performed in a meaningful way. In this thesis we have utilized biophysical analysis
tools such as EPR spectroscopy, CD, TEM and kinetics analyses in order to identify and target
conformations of the diabetic amyloid protein IAPP in the hope that these efforts will pave the
way for the development of the first successful therapeutics against T2DM.
Efforts to understand the misfolding of IAPP have led to the development of detailed structures
for the fibrillar amyloid assembly of IAPP as well as the helical, membrane-bound form of IAPP,
but the full relevance of the helical form to the disease state is not well understood (1–3). In the
second and third chapters of the work presented in this thesis we have investigated IAPP structural
transitions that occur during the misfolding of IAPP in disease-relevant settings. We conclude that
multiple disease-related lipid and lipid like molecules induce IAPP misfolding by a shared,
membrane-mediated mechanism. In addition, we have characterized the interaction between IAPP
176

and two endogenous mitochondrially-derived peptides, HNG and SHLP2. Using biophysical
techniques to follow the structural changes occurring during the misfolding of IAPP, we conclude
that both MDPs inhibit IAPP misfolding via direct interactions with misfolded amyloid seeds
(Chapter 4).
As discussed in Chapter 1, virtually all levels of the misfolding process present conformers of
IAPP that might be targeted to prevent misfolding. Unfortunately, the paucity of well-defined
structures has hindered progress towards the development of pharmacological inhibitors.
Naturally, this has made it vitally important to identify and characterize structures of IAPP that are
targetable by pharmaceuticals. In chapters 2 and 3 we identify the membrane-bound α-helix as a
key intermediate in IAPP misfolding that ties together multiple T2DM risk factors and disease
mechanisms. Obesity, dyslipidemia and insulin resistance create a β-cell environment rich in fatty
acids. High glucose levels (hyperglycemia) promote the formation of phosphatidic acid within β-
cells. Monophthalate esters are produced in the body following exposure to various plastics. These
families of molecules are associated with an increased risk of developing T2DM. By studying their
interaction with IAPP, we have found that they drive IAPP misfolding by a shared mechanism
involving the formation of a membrane-bound helical intermediate that precedes the β-sheet forms
associated with toxicity (Chapter 2). The mitochondrial lipid, cardiolipin, also enables IAPP to
form a helix in the membrane (Chapter 3). The generic nature of this mechanism and its
associations with multiple T2DM risk factors highlights the importance of the helical intermediate
in disease pathogenesis. Interventions that target and stabilize the helical conformation of IAPP
have the potential to simultaneously mitigate the influence of multiple major risk factors like
obesity, dyslipidemia, insulin resistance, and plastics exposure. Furthermore, as mitochondrial
involvement in T2DM is well-documented but poorly understood, and IAPP can be found in the
177

mitochondria (3), our findings help to establish a mechanistic connection between mitochondrial
dysfunction in T2DM and IAPP misfolding. Once again, we predict that inhibitors that can
stabilize or sequester helical IAPP intermediates will be able to intervene to help protect the
mitochondria from IAPP-mediated insult.
Other important targets exist beyond the helical intermediate. The IAPP misfolding cascade is
initiated through the spontaneous formation of misfolded IAPP conformers that seed the
propagation of the misfolded template throughout the naïve IAPP population. No reliable structural
determinants have been identified between the helix and fibril as these conformers are transient,
proceeding down the free-energy landscape smoothly. This significantly complicates the task of
creating misfolding inhibitors against these seeding conformers. How can one develop agents that
target a conformation of misfolded amyloid without first knowing the structure of the conformer?
Although our understanding of the structural elements of amyloid misfolding is relatively new, the
problem of protein misfolding itself is not. Therefore, it is reasonable to look to nature to provide
a solution where one is not readily apparent via synthetic means. This thought has led to the search
for naturally occurring misfolding inhibitors and the discovery that chaperones bind and in some
cases refold misfolded amyloids (4–6). In chapter 4 we used TEM, EPR spectroscopy, CD and
kinetics analyses to investigate the chaperone-like activity of members of a newly discovered
family of peptides, mitochondrially-derived peptides, MDPs. Our conclusion that MDPs have
chaperone-like function comes from a constellation of data that shows direct interaction between
misfolded IAPP seeds and the MDPs, the ability to inhibit misfolding, prevent propagation of
misfolded templates and no evidence of interactions between the MDPs and monomeric IAPP
(Chapter 4). Our work, however, directly tested only two MDPs, HNG and SHLP2. There are at
least six known members of the SHLP family (SHLP1-6) (7), and at least one other MDP, MOTS-
178

c (8,9) that have been described. Additional research will be required to determine the extent and
nature of MDP chaperone function in mammalian cells. Our work raises the possibility that MDPs
with chaperone-like abilities might be a part of an ancient mitochondrial defense system that can
be used to help reduce the burden caused by protein-misfolding. If so, MDPs represent an avenue
of study that might lead to a family of chaperone-like misfolding inhibitors with activity against
multiple misfolded protein species in multiple protein-misfolding diseases.
With respect to protein misfolding diseases, we live in unprecedented times. Never has the need
for avenues of treatment been greater, yet never have we known more about these diseases, nor
had such an array of tools at our disposal to deepen that knowledge. We are at a tipping point, but
have only just begun to scratch the surface of what rational intervention into the protein-misfolding
disease process can accomplish. By pursuing this avenue of research, we may find that what is
currently a family of incurable diseases becomes a family of predictable and preventable diseases.
In order to achieve this outcome, a confluence of efforts must answer certain very specific needs
and questions addressed below.
In most cases, clinical manifestation of amyloid diseases doesn’t take place until significant
damage has already been done. Without early detection, interventions to stop disease progression
will reach a glass ceiling, leaving patients permanently damaged because treatments began too
late. Thus, it is fundamentally necessary to pursue early detection of aggregation events in order
to mitigate the impact of toxic-misfolding during the sub-clinical phases of disease progression.
As we continue to gather data with today’s biophysical techniques, we must also endeavor to
improve upon those techniques. Specifically, we need to identify more non-toxic as well as toxic
pathological conformations of amyloids to use as pharmacological targets. Secondly, we must
179

find methods that to allow us to probe the events that unfold during misfolding in situ at the same
level of detail that we have been able to achieve in vitro. To do so will allow us to better understand
the relationships between specific cellular environments and protein form and function in vivo.
This will help us to differentiate between clinically-relevant disease mechanisms and mechanisms
that only apply in the test tube.
At the same time there remain many unanswered questions regarding the mechanisms already at
work in cells preventing the deleterious consequences of aberrant protein misfolding (10). By
elucidating the elements of protein homeostatic mechanisms such as autophagy, the ubiquitin
proteasome system and molecular chaperones, we will be better able to work with and enhance the
tools already available in cells to control protein-misfolding diseases.
Lastly, we have only lightly touched on the issue of amyloid transmissibility. It remains, however,
an important, yet poorly understood field. Transmission of amyloid diseases has been seen in the
past, in particular with CJD (11). As transmission of other amyloids has been seen in the laboratory
setting (11), it remains an open question as to which amyloid pathogens can be transmitted human-
to-human via transfusions, exposure to bodily fluids, or other means.

REFERENCES
1. Bedrood S, Li Y, Isas JM, Hegde BG, Baxa U, Haworth IS, et al. Fibril Structure of
Human Islet Amyloid Polypeptide. J Biol Chem. 2012 Feb 17;287(8):5235–41.
180

2. Apostolidou M, Jayasinghe SA, Langen R. Structure of alpha-helical membrane-bound
human islet amyloid polypeptide and its implications for membrane-mediated misfolding.
J Biol Chem. 2008 Jun 20;283(25):17205–10.
3. Kegulian NC, Sankhagowit S, Apostolidou M, Jayasinghe SA, Malmstadt N, Butler PC, et
al. Membrane Curvature-sensing and Curvature-inducing Activity of Islet Amyloid
Polypeptide and Its Implications for Membrane Disruption. J Biol Chem. 2015 Oct
23;290(43):25782–93.
4. Blair LJ, Sabbagh JJ, Dickey CA. Targeting Hsp90 and its co-chaperones to treat
Alzheimer’s disease. Expert Opin Ther Targets. 2014 Oct;18(10):1219–32.
5. Chaari A, Hoarau-Véchot J, Ladjimi M. Applying chaperones to protein-misfolding
disorders: molecular chaperones against α-synuclein in Parkinson’s disease. Int J Biol
Macromol. 2013 Sep;60:196–205.
6. Arawaka S, Machiya Y, Kato T. Heat Shock Proteins as Suppressors of Accumulation of
Toxic Prefibrillar Intermediates and Misfolded Proteins in Neurodegenerative Diseases.
Curr Pharm Biotechnol. 2010;11(2):158–66.
7. Cobb LJ, Lee C, Xiao J, Yen K, Wong RG, Nakamura HK, et al. Naturally occurring
mitochondrial-derived peptides are age-dependent regulators of apoptosis, insulin
sensitivity, and inflammatory markers. Aging. 2016 Apr 10;
8. Fuku N, Pareja-Galeano H, Zempo H, Alis R, Arai Y, Lucia A, et al. The mitochondrial-
derived peptide MOTS-c: a player in exceptional longevity? Aging Cell. 2015
Dec;14(6):921–3.
181

9. Lee C, Zeng J, Drew BG, Sallam T, Martin-Montalvo A, Wan J, et al. The mitochondrial-
derived peptide MOTS-c promotes metabolic homeostasis and reduces obesity and insulin
resistance. Cell Metab. 2015 Mar 3;21(3):443–54.
10. Hipp MS, Park S-H, Hartl FU. Proteostasis impairment in protein-misfolding and -
aggregation diseases. Trends Cell Biol. 2014 Sep;24(9):506–14.
11. Prusiner SB. Biology and genetics of prions causing neurodegeneration. Annu Rev Genet.
2013;47:601–23.

Asset Metadata

Creator Okada, Alan Kiyoshi (author)

Core Title Enhancing and inhibiting diabetic amyloid misfolding

Contributor Electronically uploaded by the author (provenance)

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Integrative Biology of Disease

Publication Date 08/08/2016

Defense Date 07/15/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Amyloid,Diabetes,islet amyloid polypeptide,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Advisor Siemer, Ansgar (committee chair), Chen, Jeannie (committee member), Chow, Robert (committee member), Langen, Ralf (committee member)

Creator Email alan.k.okada@gmail.com,aokada@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-639571

Unique identifier UC11305857

Identifier etd-OkadaAlanK-4765.pdf (filename),usctheses-c3-639571 (legacy record id)

Legacy Identifier etd-OkadaAlanK-4765.pdf

Dmrecord 639571

Document Type Dissertation

Format application/pdf (imt)

Rights Okada, Alan Kiyoshi

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Abstract (if available)

Abstract The misfolding and aggregation of proteins is associated with some of the most devastating diseases of the modern era, including type 2 diabetes mellitus, Alzheimer disease and Parkinson disease. In type 2 diabetes mellitus (T2DM), the misfolding of the 37-amino acid peptide, islet amyloid polypeptide (IAPP), is associated with cytotoxicity of insulin producing β-cells. It is thought that efforts to intervene in the misfolding process will lead to the development of therapeutics for diseases such as T2DM. To facilitate these efforts, it is important to identify conformations of proteins involved in disease-associated misfolding cascades. In this thesis we investigate the misfolding of IAPP using a combination of thioflavin T fluorescence (ThT), circular dichroism (CD), electron paramagnetic resonance spectroscopy (EPR), and transmission electron microscopy (TEM). We find that liposomes loaded with anionic lipids and lipid-like molecules associated with T2DM risk factors (e.g. obesity and plastics exposure), such as phosphatidic acid (PA), oleic acid (OA), a fatty acid, and monobenzylester phthalate (MBzP), a monophthalate ester, dramatically enhance the misfolding of IAPP. Such membrane interactions with IAPP lead to the rapid formation of a membrane-bound and metastable α-helical intermediate that slowly transitions into fibrillar amyloid on the surface of the membrane. Due to the mitochondrial dysfunction seen in T2DM, we also investigated the interaction between the negatively charged cardiolipin (CL), a uniquely mitochondrial lipid and IAPP. We found that CL enhances the misfolding of IAPP in a manner similar to that of PA, OA and MBzP. These data are consistent with the hypothesis that the ability to induce membrane-mediated misfolding of IAPP is a generic trait of negatively charged membranes. This highlights the importance of the α-helical conformation of IAPP as a drug development target since multiple risk factors for T2DM, as well as mitochondria-like membranes, induce misfolding of IAPP through the same α-helical conformer. ❧ We then investigated the ability of two mitochondrial derived peptides (MDPs) to alter the course of IAPP misfolding and found that both MDPs inhibit misfolding by direct, chaperone-like interactions with misfolded amyloid seeds. This chaperone-like activity makes these MDPs exciting new prospects for development as T2DM therapeutics since they provide the opportunity to target key conformations along the misfolding pathway without necessitating atomistic level characterization of the conformers themselves. ❧ Taken together, this work focuses on two key targets along the IAPP misfolding pathway. First, we highlight the importance of the α-helical intermediate of IAPP formed during membrane-mediated misfolding. Second, we characterize two naturally occurring peptides of the MDP family with chaperone-like activities that target misfolded IAPP conformations to prevent propagation by misfolded templates. These studies are a part of larger, ongoing efforts to understand the amyloid misfolding process and use that knowledge to develop therapeutics that will one day lead to a cure for the millions of people suffering from T2DM and other protein misfolding diseases.