What can Drosophila tell usabout serpins, thrombosisand dementia?Robin Carrell1* and Javier Corral2

SummaryThe validity of the fruit-fly as a model of human diseasehas been confirmed in a striking way by Green andcolleagues.(1) They show that the mutations causing anecrotic disease phenotype in Drosophila, preciselymirror those resulting in a group of well-studied but per-plexing diseases in the human. These diseases, rangingfrom thrombosis to dementia, arise from mutationscausing a conformational instability of serpin proteaseinhibitors. The findings provide clues as to the unusualseverity and variable onset of such conformational dis-eases and demonstrate the potential of Drosophila as amodel for their future study. BioEssays 26:1–5, 2004.� 2003 Wiley Periodicals, Inc.

Introduction

The best of all experimental animals is man himself. The

experiments of nature resulting in genetic disease are not

limited by our own preconceptions, so they consistently open

new and unexpected understandings, not only of pathology

but also of basic biology. A classical example is the common

inherited deficiency of the main protease inhibitor in plasma,

a-1-antitrypsin (antitrypsin). As its name indicates, antitrypsin

was originally thought to function as an inhibitor of trypsin but

the recognition that its deficiency predisposed to the destruc-

tive lung disease emphysema(2) revealed its true action, as an

inhibitor of the elastase released by inflammatory white cells.

This in itself opened the wider concept of the balance between

proteases and inhibitors that is central to the health of higher

organisms. However, the real bonus from the discovery of this

genetic abnormality was the realisation that the mutation

responsible for the deficiency did not limit the synthesis of

antitrypsin but rather caused its aggregation in the hepato-

cytes of the liver. The consequence of this intracellular protein

aggregation is a progressive loss of hepatocytes leading

eventually to the degenerative liver disease cirrhosis.(3) The

special significance of this finding is that antitrypsin belongs to

a family of serine protease inhibitors, the serpins, which have a

tightly conserved and well-characterised structure.(4) We can

now see in crystallographic detail the way in which mutations

can cause a change in the serpin fold, with resultant inter-

molecular linkage and polymer formation. Although these

mutations were originally identified in antitrypsin, precisely

homologous mutations have now been identified in other

human serpins. In antithrombin, the principal inhibitor of

coagulation, the same mutations result in thrombosis, in C1-

inhibitor, they can cause fatal immune sensitivity and, most

significantly, in the brain-specific neuroserpin, the mutations

result in a typical late-onset dementia. What is remarkable is

the way that all the findings from these diverse serpins fit

coherently together to provide a detailed prototype for what are

now known as the conformational diseases.(5)

Drosophila transgenics

The human model cannot of course be used for elective

experimentation and in its stead attention has switched to the

transgenic mouse. But the mouse too has its limitations: in

complexity, in generation time and in the expense of breed-

ing and upkeep. Recently however, a new model for human

disease has appeared, or rather an old model has found a new

role. The fruit-fly Drosophila, has well-studied genetics, there

is no need for an animal house—a grapefruit and a bell jar will

suffice—and at the end of the experiment the colony can be

homogenised and run down a column to isolate the molecules

of interest. It seems ideal! And so it is proving to be, with a

series of experiments from different groups now providing

insights into some of the most puzzling of all disorders, the

slow-onset dementias.(6–9) In this way, Drosophila provides a

surrogate for the study of the effects of genetic aberrations, but

the approaches used to study diseases such as Huntington’s

and Parkinson’s involve the overexpression of transgenic

proteins at a selected tissue level. Can the results from such

specific targeting be extrapolated to the whole-organism

consequences of human disease?

Validation of Drosophila as a whole-organism model of

human disease comes from the recent identification by Green

and colleagues(1) of a phenotype that mimics in detail the

BioEssays 26:1–5, � 2003 Wiley Periodicals, Inc. BioEssays 26.1 1

1University of Cambridge, Cambridge Institute of Medical Research,

Cambridge, UK.2University of Murcia, Centro Regional de Hemodonacion, Murcia,

Spain.

*Correspondence to: Robin Carrell, University of Cambridge, Cam-

bridge Institute of Medical Research, Cambridge, UK CB2 2XY, UK.

E-mail: rwc1000@cam.ac.uk

DOI 10.1002/bies.10407

Published online in Wiley InterScience (www.interscience.wiley.com).

What the papers say

diseases that arise from serpin mutations in man. The nec

(necrotic) phenotype in Drosophila is characterised by the

development of patches of epithelial necrosis followed by early

death. The phenotype arises from mutations in the gene

coding for the Nec protein, a protease inhibitor that controls the

Toll-mediated immune response of Drosophila.(10,11) The Nec

inhibitor is a typical serpin, with a core structure that is closely

homologous to that of antitrypsin and which is similarly syn-

thesised in the insect equivalent of the liver, the fat body. The

necrotic phenotype results from a loss of the inhibitory function

of the Nec serpin, so the predominant mutations would be

expected to be those causing a failure of expression. But

the surprise finding of Green et al. was that the majority of the

identified mutations resulted in the substitution of residues

known to be critical to the conformational stability of serpins. In

particular, two of the mutations, at Glu342 (serpin template

numbering, Ref. 4) were identical to the destabilising mutation

Glu342Lys causing the common form of inherited a-1-

antitrypsin deficiency.(2,12) This identity of mutations resulting

in defined pathologies provides a striking endorsement of

Drosophila as a valid model of human disease. Moreover, the

overall pattern of mutations causing the necrotic phenotype,

which indicate that a prime contributor to the pathology is

the conformational instability of the Nec serpin rather than its

mere deficiency, has immediate implications for the human

serpinopathies, even beyond those recognised by Green and

colleagues.

Serpin mobility and disease

The serpins are molecular machines. The key to their function

and to their evolutionary success is their mobile mechanism.

The many other families of serine protease inhibitors function

in a simple way by extending a substrate-like reactive centre

peptide loop that passively blocks the active centre of the

target protease.(13) The serpins too have a similarly configured

and exposed reactive loop that acts as a protease-specific

bait. But when the reactive centre loop is cleaved by the

protease, the serpins undergo a unique and extraordinary

conformational change (Fig. 1A). The cleaved peptide loop

moves to fill the middle strand of the main b-sheet of the

molecule, and in doing so carries with it the entrapped pro-

tease. The protease is flung a distance of 70 A with an

accompanying plucking and distortion of its structure that

ensures its destruction.(14) It is this ability of the serpins to give

effectively irreversible inhibition that has led to their selection

as the controlling inhibitors of the proteolytic pathways of the

cell and higher organisms. The subtleties of these control

processes is evident in the coagulation pathways of man

where serpins—antithrombin, heparin cofactor II, and protein

C inhibitor—control the formation of the fibrin clot, whereas

other serpins—antiplasmin and the plasminogen activator

inhibitors—control its dissolution. Abnormalities in the function

of any of these serpins lead to a shift in this delicate balance,

with the consequent onset of either thrombosis or haemor-

rhage. This disease association has stimulated a series of

crystallographic studies that makes the serpins among the

best understood, in structural terms, of all protein families.

There are now some 30 crystallographic structures of different

conformations of serpins, which together provide accurate

video depictions of the way the molecule can change its fold.

The crystallographic videos show not only the co-ordinated

changes in fold of the serpin skeleton, but also the precise

interactions and movements of individual amino acid side-

chains. These interactions are critical to the efficiency of

the hinges of the reactive centre loop, and of the central shutter

region of the mainb-sheet (Fig. 1B) where strands 3 and 5 slide

apart to allow entry of the loop as strand 4. Evidence for the

tight specificity of these sidechain interactions came from a

study ten years ago of some 100 known dysfunctional mutat-

ions in serpins.(15) The conclusion from this, that mutations

causing disease would mainly involve the hinges of the re-

active loop and the shutter region at the focus of the opening of

the sheet, has been borne out in the many other mutations that

have since been identified in serpins. This is highlighted by the

subsequent finding of the same pattern in recently recognised

serpins, notably in the mutations of neuroserpin that result in

encephalopathies and dementia(16,17) and, as discussed here,

in the mutations of the Drosophila serpin that result in the nec

phenotype.(1)

Conformational disease, thrombosis

and dementia

The bonus from the nec Drosophila findings is the additional

insights that they provide into the dysfunctions of the serpins

with relevance to the mechanisms of the conformational

diseases as a whole. The conformational diseases comprise

one of the most perplexing and arcane group of disorders in

medicine.(5) They have diverse manifestations, often with slow

onset, as in the common dementias and the spongiform

encephalopathies, and sometimes with a sudden severe onset

in later life, as in episodic familial thrombosis. A shared feature,

however, of all the conformational diseases is that they result

from a change in fold or size of a constituent protein, with

associated intermolecular b-linkage and, characteristically,

with tissue deposition. These are the changes that are typified

by the common (Glu342Lys) antitrypsin deficiency in man, and

the focus of the paper of Green and colleagues is under-

standably on their finding of the same mutation in the nec

serpin of Drosophila. This substitution of the conserved

glutamate, at the hinge of the reactive loop of the serpin, will

have direct consequences in predictably slowing entry of the

loop into the sheet with a resultant decrease in inhibitory

activity. But a more serious conformational consequence of

this hinge dysfunction is that it will allow partial entry of the loop

into the sheet with an accompanying opening of the lower half

of the sheet (Fig. 1B). This opening of the sheet potentially

What the papers say

2 BioEssays 26.1

allows the insertion of the loop of another molecule to give

intermolecular loop-sheet linkages. Such sequential loop-

sheet linkage, with the formation of long polymers,(18) has

been well characterised with the Glu342Lys mutant of a-1-

antitrypsin. It is this formation of polymers that gives the ‘‘gain-

of-function’’ disadvantage to the mutant antitrypsin that is

another characteristic feature of the conformational diseases.

In other words, the conformational instability confers a

disadvantage over and beyond that of loss of activity. Thus

the loss of inhibitory protection due to the plasma deficiency of

Figure 1. Serpin conformational mobility A: Inhibition by the serpin occurs when the protease (cyan) cleaves its reactive centre loop

(yellow) which then inserts into the middle, strand 4, position of the A sheet (red). The distortion that accompanies this displacement of the

protease results in a gross disordering of its structure (ghost cyan).(14) B(i): Insertion of the reactive loop in the intact molecule is limited by

interactions in the shutter region underlying the bifurcation of the sheet (encircled). ii: Mutations in this or other hinge areas allow the full

opening of the sheet with, as arrowed, the potential entry of the reactive loop of another molecule. iii: Giving sequentially the formation of

loop-sheet polymers.

What the papers say

BioEssays 26.1 3

a-1-antitrypsin, as would also occur with a null variant, explains

the destructive lung disease associated with the deficiency of

the inhibitor. But the gain-of-function disadvantage is due to

the intracellular aggregation of the polymerised antitrypsin that

eventually leads to liver cirrhosis.

Is there a gain-of-function disadvantage of the nec

mutations? At first sight it would seem not, as the necrotic

phenotype can similarly result from uncomplicated null mutat-

ions. Moreover, the Nec serpin forms shorter oligomers rather

than the long polymers that result in the intracellular aggrega-

tion associated with the loss of hepatocytes in the liver with

mutant a-1-antitrypsin, and with the loss of neurones with

homologous mutants of neuroserpin. The latter example is of

importance. Four mutations in the mobile shutter region of

neuroserpin have each been identified as causing progres-

sive neurodegeneration.(17) The findings, at both cellular and

molecular levels, closely mimic those seen with antitrypsin,

with the formation of large neuronal inclusions of the poly-

merised neuroserpin. The clear relationship between the

magnitude of the inclusions, the destabilising effect of each

mutation, and the age of onset of the dementia, provides

strong evidence for the direct toxicity of the intracellular

aggregation of the mutant serpin. The nec mutations are

homologous with mutations in both neuroserpin and in a-1-

antitrypsin (Fig. 2), but it is clear that the consequences of

the mutations in the Nec serpin are distinctly different. These

consequences more closely resemble those seen with homo-

logous mutations in antithrombin, and thus the findings with

the nec Drosophila provide an unexpected insight into the

changes underlying thrombotic disease.

This variation in the disease consequences is well illu-

strated by the nec mutation identified in two of the affected

alleles, at Gly 386, a conserved residue that is essential to the

tight packing and hence to the function of the shutter mech-

anism. Mutation of Gly386 in neuroserpin results in the

formation of long-chain polymers and the onset of severe

Figure 2. Mutations in Drosophila causing the necrotic disease phenotype are seen to affect conserved or invariant residues known to

be critical for conformational stability. Most of these sites are identical to the sites of mutations in human serpins known to cause a variety

of diseases as indicated.

What the papers say

4 BioEssays 26.1

neurodegeneration by age 12. Whereas in antithrombin, we

find in our recent unpublished work, that the mutation of the

same glycine results primarily in the formation of dimers and a

predispositon to unexpectedly severe thrombosis in the adult.

So the results with antithrombin closely parallel the findings

with the two mutants at Gly 386 identified in the nec

Drosophila. The shared features are oligomer formation with

an exacerbation of the consequences of deficiency, rather

than the polymer formation with massive intracellular aggre-

gation observed with a Gly 386 mutation in neuroserpin. The

comparison becomes even more pointed with the finding in

humans(19) of an individual with a homozygous Glu342Lys

mutation in the anticoagulant serpin, heparin cofactor II. This

results in a deficiency of the inhibitor, which exacerbates an

existing predisposition to thrombosis. Again, the pathology

follows the pattern observed with the same mutation in the Nec

serpin. The feature in common in both the Drosophila necrotic

and the human thrombotic is not just the same underlying

mutations but also a severity of consequences, beyond that

of mere deficiency, that implies the presence of an as yet

unidentified gain-of-function disadvantage.

The great plus for the Drosophila system, as compared to

the human studies, is that Green and colleagues have been

able to directly demonstrate the disadvantageous effect of the

destabilising mutations by expression of the mutant serpin

superimposed on that of the normal wild-type alleles. Their

results clearly show a gain-of-function disadvantage, and fit

with other recent work on the inhibitors of coagulation in

opening prospects of an experimental lead to the still unknown

factors that contribute to sudden and severe thrombosis. One

suspected factor is the small increases in temperature that

accompany incidental infections. There is compelling clinical

evidence, in families with conformationally unstable mutants of

antithrombin such as the Gly386 variant, of the precipitation by

infections of the sudden onset of thrombosis. By inference

then, infections with fever will similarly be expected to exacer-

bate the cumulative damage that occurs with homologous

mutants of a-1-antitrypsin and neuroserpin and hence accel-

erate the onset of the liver disease and dementia. It is a critical

but controversial conclusion that has direct relevance to the

clinical management not only of the serpinopathies but also of

other more common conformational dementias and diseases.

The problem in settling this controversy has been that,

although the small changes in body temperature that occur

with infections can be shown to cause the aggregation of the

unstable serpins in vitro, the in vivo observations are anecdotal

rather than experimental. The special value of the Drosophila-

based model is that it allows in vivo experimentation, as Green

and colleagues show in their paper by the demonstration of a

decrease in survival time of nec Drosophila at increased

temperatures.

References1. Green C, Brown G, Dafforn TR, Reichhart JM, Morley T, Lomas DA, Gubb

D. Drosophila necrotic mutations mirror disease-associated variants of

human serpins. Development 2003;130:1473–1478.

2. Laurell C-B, Eriksson S. The electrophoretic a1-globulin pattern of serum

in a1-antitrypsin deficiency. Scand J Clin Lab Invest 1963;15:132–140.

3. Eriksson S, Carlson J, Velez R. Risk of cirrhosis and primary liver cancer

in alpha1-antitrypsin deficiency. N Eng J Med 1986;314:736–739.

4. Huber R, Carrell RW. Implications of the three-dimensional structure of

a1-antitrypsin for structure and function of serpins. Biochemistry 1989;28:

8951–8966.

5. Lomas DA, Carrell RW. Serpinopathies and the conformational demen-

tias. Nature Review Genetics 2002;3:759–768.

6. Jackson GR, Salecker I, Dong X, Yao X, Arnheim N, Faber PW,

MacDonald ME, Zipursky SL. Polyglutamine-expanded human huntingtin

transgenes induce degeneration of Drosophila photoreceptor neurons.

Neuron 1998;21:633–642.

7. Warrick JM, Paulson HL, Gray-Board GL, Bui QT, Fischbeck KH,

Pittman RN, Bonini NM. Expanded polyglutamine protein forms nuclear

inclusions and causes neural degeneration in Drosophila. Cell 1998;93:

939–949.

8. Feany MB, Bender WW. A Drosophila model of Parkinson’s disease.

Nature 2000;404:394–398.

9. Marsh JL, Walker H, Theisen H, Zhu YZ, Fielder T, Purcell J,

Thompson LM. Expanded polyglutamine peptides alone are intrinsically

cytotoxic and cause neurodegeneration in Drosophila. Hum Mol Genet

2000;9:13–25.

10. Levashina EA, Langley E, Green C, Gubb D, Ashburner M, Hoffmann JA,

Reichhart JM. Constitutive activation of toll-mediated antifungal defense

in serpin-deficient Drosophila. Science 1999;285:1917–1919.

11. Green C, Levashina E, McKimmie C, Dafforn T, Reichhart JM,

Gubb D. The necrotic gene in Drosophila corresponds to one of a

cluster of three serpin transcripts mapping at 43A1.2. Genetics 2000;

156:1117–1127.

12. Jeppsson J-O. Amino acid substitution Glu!Lys in a1-antitrypsin PiZ.

FEBS Lett 1976;65:195–197.

13. Bode W, Huber R. Structural basis of the endoproteinase-protein

inhibitor interaction. Biochim Biophys Acta 2000;1477:241–252.

14. Huntington JA, Read RJ, Carrell RW. Structure of a serpin-protease

complex shows inhibition by deformation. Nature 2000;407:923–926.

15. Stein PE, Carrell RW. What do dysfunctional serpins tell us about

molecular mobility and disease? Nature Struct Biol 1995;2:96–113.

16. Davis RL, et al. Familial dementia caused by polymerisation of mutant

neuroserpin. Nature 1999;401:376–379.

17. Davis RL, et al. Association between conformational mutations in

neuroserpin and onset and severity of dementia. Lancet 2002;359:

2242–2247.

18. Sivasothy P, Dafforn TR, Gettins PG, Lomas DA. Pathogenic alpha 1-

antitrypsin polymers are formed by reactive loop-beta-sheet A linkage.

J Biol Chem 2000;275:33663–33668.

19. Villa P, Aznar J, Vaya A, Espana F, Ferrando F, Mira Y, Estelles A.

Hereditary homozygous heparin cofactor II deficiency and the risk of

developing venous thrombosis. Thromb Haemost 1999;82:1011–1014.

What the papers say

BioEssays 26.1 5