A R T I S T ’ S N O T E

Protein Sculptures: Life’s Building Blocks Inspire Art

Julian Voss-Andreae

There is in Nature a limitless variety of shapes and rhythms (andthe telescope and microscope have enlarged the field) from whichthe sculptor can enlarge his form-knowledge experience.

—-Henry Moore [1]

In his 1968 book Beyond Modern Sculpture, the visionary authorJack Burnham asserted, “In a very real sense [D’Arcy W.Thompson’s book] On Growth and Form [2] stands on thethreshold between that world of natural forms which is still ac-cessible to the sculptor, and the world of molecular bonds andprotein chains completely out of his reach” [3].

What Burnham did not take into account at the time of thisassessment was the rapid advancement of technology thatwould one day provide artists with the tools necessary to gaininspiration from nature beyond what can be seen by the unaided eye. One generation after his writing, powerful com-puters came into widespread use. At the same time, experi-mental techniques came to routinely resolve structures on asubatomic scale, and laboratories started making their exper-imental results accessible on the Internet. Today anyone witha computer can, in principle, download tens of thousands ofdifferent protein structures [4] and create highly illusionisticthree-dimensional renditions of these data.

PROTEINSBoth life and inanimate matter consist of atoms. Atoms are thebuilding blocks from which all the matter that surrounds usand all the matter we experience as “ourselves” is made. Atomscombine to form molecules. Molecules encountered in bio-chemistry differ considerably, for the most part, from mole-cules found in inanimate matter. The fundamental moleculeof life is DNA, which contains the “blueprint” for each or-ganism’s form as well as for its function. The double helix,with its sequence of base pairs, is essentially a one-dimensional[5] strand of information. Every triplet of base pairs in a genecodes for one of the 20 different amino acids that are the build-ing blocks of life. Proteins, chains of linked amino acids, arethe next level of important building blocks. The physical prop-erties of the different amino acids cause them to fold and windinto well-defined 3D configurations determined by their

sequence. The process of proteinbiosynthesis and folding is the pointat which life makes the transitionfrom one-dimensional DNA intothree-dimensional bodies. Proteinsplay a key role in structure and func-tion from cell to organism. The di-verse structures of proteins give riseto a stunning variety of functions[6]. Enzymes, an important class ofproteins, are catalysts needed to reg-ulate all biochemical reactions. Invertebrates the antibodies form a major line of defense againstforeign organisms. Our every movement results from the con-traction and relaxation of muscles, resulting in turn from the

© 2005 ISAST LEONARDO, Vol. 38, No. 1, pp. 41–45, 2005 41

Julian Voss-Andreae (artist/scientist), 2146 NE 10th Avenue, Portland, OR 97212, U.S.A. E-mail: <JulianVossAndreae@hotmail.com>. Web site: <www.JulianVossAndreae.com>.

A B S T R A C T

The author takes a literal lookat the foundation of our physicalexistence by creating sculpturesof proteins, the universal partsof the machinery of life. For him,it is less important to copy amolecule accurately in all itsdetails than to find a guidingprinciple and follow it to seewhether it yields artisticallyinteresting results. The mainidea underlying these sculpturesis the analogy between thetechnique of mitered cuts andprotein folding. The sculpturesoffer a sensual experience of aworld that is usually accessibleonly through the intellect.

Fig. 1. The principle of mitered cuts. (© Julian Voss-Andreae) A picture frame is constructed by cutting a one-dimensional piece of wood (a) at 45˚ (b), flipping every other piece (c), and reassembling the pieces in the same order (d). Mitered cuts can be applied such that the material occupies all three dimensions after reassembly (e and f).

ART

AND

BIO

LOG

Y

delicately regulated sliding of specificproteins in muscle cells past one another.Still other proteins act as adjustable chan-nels through which ions are passed fromone side of a membrane to the other, re-sulting in phenomena such as the trans-mission of electric signals along nerves,which is the physical basis of our thoughtsand our senses.

PRINCIPLEOne of the fundamental properties ofproteins is that they are 3D objects de-spite their essentially one-dimensionalstructure. We are familiar with similarphenomena in our everyday lives. An ex-ample is a piece of wire bent many timesinto a compact 3D shape. A rigid mate-rial like a piece of lumber, however, canalso be “folded” by applying mitered cutsand reassembling it. For instance, an ob-ject such as a picture frame is constructedin this fashion by cutting a long piece ofwood four times at 45˚ (see Fig. 1a–d).The pieces are glued together in thesame order after every other piece is ro-tated 180˚ around the length of the wood,which results in a “folding” of 90˚—thatis, twice the cutting angle. The pictureframe example is still only 2D, but it isalso possible to produce an object occu-pying all three dimensions by cutting atdifferent angles (see Fig. 1e–f). The pieceso reassembled is essentially the same as

and the material is cut along the linesand reassembled.

A protein structure database [8] pro-vides my program with the position ofeach amino acid. The amino acids in theprotein form a chain of identical flat rec-tangular units called peptide units [9].Particular carbon atoms, usually denotedwith the subscript α, connect these units.The amino acids do not differ in the pep-tide units making up the backbone of theprotein. The difference lies only in theside chains departing from the Cα atoms,whose physical properties are the keyforce guiding protein folding. I use thepositions of the Cα atoms [10] to computecutting instructions. That means thateach piece of the sculpture, extendingfrom the center of one joint to the next,corresponds to one peptide unit in themolecule.

SCULPTURESMy first sculpture was a small (58-amino-acid) protein called Bovine PancreaticTrypsin Inhibitor (BPTI) [1bpi] [11], aprotein that inhibits the digestive enzymetrypsin in cows. I first came across BPTIin a physics journal [12], which showedthe protein using different kinds of mod-els (see Color Plate D No. 2,a–d), illus-trating the problem one encounters invisualizing these structures. Color PlateD No. 2,a shows all atoms rendered asspheres. Color Plate D No. 2,b and ColorPlate D No. 2,c emphasize the chemical

before, because no material was added or lost. It extends into three dimensionsonly by virtue of a rearrangement of itsparts. The principle of creating sculp-tures through the technique of miteredcuts lends itself very well to representingprotein structures, because both sharethe fundamental property of being one-dimensional objects occupying three di-mensions. It is surprisingly difficult tomake miter-cut sculptures using a pre-conceived plan, but the well-defined geo-metric properties of miter-cut sculpturesallow for a computational treatment ofthe problem.

REALIZATIONI developed a computer program that al-lows me to turn any given sequence of 3Dpoints into a miter-cut sculpture [7]. Thesoftware renders the sculptures realisti-cally from any point of view, allowing meto rotate the virtual object at will and lookat it as if it were in real space (see ColorPlate D No. 2,e and Fig. 2). My programalso provides me with detailed cutting in-structions by computing all angles andlengths needed for a physical realizationof the sculpture. A list containing the po-sitions of the four points of intersectionbetween each cutting plane and the ma-terial’s edges is generated. This allows foran easy marking of the correct distancesalong the edges. The marks are then con-nected (with the additional control of anangle measurement using a protractor),

42 Voss-Andreae, Protein Sculptures

ART

AND

BIO

LOG

Y

Fig. 2. Virtual model for a sculpture of the Green Fluorescent Protein GFP[1emg]. (© Julian Voss-Andreae) This protein makes a certain jellyfish glow green in the dark and is used extensively in biological research. Its structure, a bird-cage-like barrel made up of 11 bars spiralingup and down the surface of a cylinder, isextraordinarily beautiful. The peptide chainis rendered in different shades of gray to aid in the recognition of the complicatedstructure.

Fig. 3. Kalata, painted steel, length 3 ft, 2002. (© Julian Voss-Andreae) The sculpture waswelded from 2-×-2-in square steel tubing and spray-painted in blood red. The uteroactiveprotein Kalata B1 [1k48] has been isolated from traditional African herbal medicine, where it is used to accelerate labor in childbirth. The cyclic structure consists of 29 amino acids and has the same topology as the Möbius strip.

structure of the molecule. The path ofthe backbone is colored from red to or-ange, yellow, green and blue, and the sidechains are gray. The rendition in ColorPlate D No. 2,d, the so-called ribbonmodel, emphasizes the backbone of theprotein and comes therefore the closestto depicting a protein as a miter-cutsculpture, shown in Color Plate D No. 2,eand Color Plate D No. 2,f. Color Plate DNo. 2,e is the virtual model of the sameprotein rendered by my computer pro-gram and displayed from the same pointof view as Color Plate D No. 2,a–d. ColorPlate D No. 2,f is a photo of the 15-in-highsculpture BPTI (2001). I chose this pro-tein to experiment with the techniqueand to get a feeling for how a virtualsculpture compares to a real one. Thematerial is wood with a cross section of 3⁄4 × 3⁄4 in. The cuts on the 9-ft-long pieceof wood were done with a handsaw, andthe pieces were glued together. Thesculpture was then spray-painted ultra-marine blue and the spiraling parts (theso-called α-helices) warm cadmium yel-low to balance the visual tension presentin the standing piece.

My next project was Green FluorescentProtein (GFP) [1emg] [13]. This proteinfrom the Pacific Northwest jellyfish Ae-quorea victoria consists of almost 240amino acids arranged in a beautiful bird-cage-like structure (see Fig. 2) [14]. GFPhad initially sparked my interest in pro-tein structure when, as a graduate stu-dent in Anton Zeilinger’s research groupin Vienna, Austria, I investigated the pos-sibility of using it to extend the demon-stration of quantum mechanical wavebehavior from Buckminsterfullerenes(C60) [15,16] to large biomolecules [17].

debates. I began two GFP wood sculp-tures, but because of the low accuracy of the power saw I used and the largenumber of amino acids, which led to anaccumulation of errors, it proved impos-sible to assemble the pieces as originallyplanned [21].

For my next material I moved on tosteel, which allows for accurately ma-chined cuts with strong welded joints. Ichose the cyclic protein Kalata B1 [1k48][22], a 29-amino-acid-long moleculefound in the African plant Oldenlandiaaffinis. Traditional African herbal medi-cine uses extracts of O. affinis to acceler-ate labor in childbirth, and Kalata B1 isthe main uteroactive agent [23]. Thesculpture shown in Fig. 3 was built from2-×-2-in square steel tubing and finishedin a glossy red, alluding to the blood ofchildbirth.

Any material where the shape of thecross-sectional area looks identical aftera rotation of 180˚ is suitable for miter-cutsculptures. Therefore materials with a cir-cular cross-section, such as tree trunks,can also be used. Wood, like steel, is amajor building block material in humanconstruction and can therefore serve inan analogy to the building-block charac-ter of proteins. For my next protein sculp-ture I used a 33-ft-long piece of a trunkof Douglas fir, the most commonly usedlumber wood in the United States [24].In addition to wanting to work on a muchlarger scale, I wanted to explore the pos-

GFP is one of the most widely used pro-teins in biological research because it isexpressed readily in many organismsafter gene transfer, allowing it to be usedas a marker of gene expression and pro-tein targeting [18]. The possibility of cre-ating animals that glow green underultraviolet light by inserting the GFPgene into their genomes [19] has at-tracted not only scientists. The creationof a GFP rabbit by Eduardo Kac as a pieceof “transgenic art” has recently surprisedthe art world [20] and triggered heated

Voss-Andreae, Protein Sculptures 43

ART

AND

BIO

LOG

Y

Fig. 4. Large Fir Alpha Helix, Douglas fir and steel, length 11 ft, 2002. (© Julian Voss-Andreae)The sculpture was made out of a 33-ft piece of a 40-year-old tree.

Fig. 5. Tall Fir Alpha Helix, Douglas fir and steel, height 10 ft, 2003. (© Julian Voss-Andreae)Unlike in the alpha helix shown in Fig. 4, the cuts in this sculpture are applied very close toone another to yield the highest possible density. The right panel shows a detail.

sibility of creating an object out of iden-tical subunits. The α-helix, which formsthe spiraling part of proteins such as theones seen in the BPTI structure (seeColor Plate D No. 2), is an ideal choice,as it is one of the most abundant struc-tural elements found in proteins. Astraight fir that had recently died wasfelled and cut into identical pieces witha chain saw. The two angles determiningeach cut were measured with custom-built devices [25]. I used three bracketswelded from ¼-in steel attached with 10 large screws for each connection be-tween the pieces to create joints strongenough to support the wood’s enormousweight. The tree was left in its originalstate, with bark and lichen on it, to en-hance the contrast between the naturaltree and its forced rearrangement (seeFig. 4).

SIGNIFICANCEThe demonstration of the in vitro syn-thesis of urea in the early 19th centuryshowed that, contrary to the leading par-adigm of that time, organic chemistry isnot confined to living organisms. Todaywe witness the isolation, transfer betweenorganisms and expression of single genesencoding particular proteins. Genetic en-gineering is, like the synthesis of urea,just another step in the merging of twotraditionally separate fields: life and atechnology based on life’s buildingblocks. Life’s building blocks are gener-ally considered inanimate and thereforeperceived as fundamentally differentfrom life. It is interesting to associate lifeand its building blocks with the poles organic and constructive, using a concep-tion from sculpture theory. HerbertRead, who was one of the major theo-reticians to promote the interpretationof the history of sculpture as an oscilla-tion between organic and constructivetendencies, made this connection ex-plicit in 1952: “We have seen that con-structive elements underlie all naturalphenomena; that organic growth followslaws, and involves structures, which areas geometrical, or mathematical, as any-thing created by a constructivist artist”[27]. My protein sculptures have aspectsof both organic and geometric objects.Like real proteins, my sculptures consistof “constructive” building blocks withvery simple geometry, but become “or-ganic” by virtue of the rearrangement ofthe parts in complex ways.

Works in both science and art embodythe most fundamental expressions of thehuman spirit. My work is an explorationand probing of the accepted divide be-tween science and art as either primarilyintellectual or primarily emotional. Thesculptures presented in this article playon the sensuality and beauty that under-lie sense and being itself.

POSTSCRIPTSince writing this article, I became awareof other sculptors who present physicalrealizations of proteins in a fine art con-text [28]. Byron Rubin, a crystallogra-pher, who invented a tool (“Byron’sBender”) for making protein modelsfrom bent wire, is a pioneer in the fieldof making protein sculptures. The smallwire structures made with his tool werethe easiest to manipulate and most por-table models available at the time. Theybecame very popular among researchersbefore computers were capable of pro-viding them with virtual models and were

For a second α-helix sculpture, a 26-ft-long dead Douglas fir was felled and cutinto over 100 pieces. This time I con-sciously diverged from making an accu-rate model of the protein element bysuccessively shortening the length of thepieces proportional to the decreasing di-ameter [26]. The pieces were then re-assembled as a vertically standing spiral.The accumulation of small errors in com-bination with the organic shape of thetree caused the piece to assume a beau-tiful organic movement with a striking re-semblance to a human spine (see Fig. 5).The emergence of such an unexpectednew level of meaning is highly welcome.In addition to its intellectual side, mywork has an equally strong intuitive andirrational side that causes my pieces tostop working as scientific models and be-come pure art objects.

44 Voss-Andreae, Protein Sculptures

ART

AND

BIO

LOG

Y

Fig. 6. Green Fluorescent Protein, steel with process marks, height 5 ft 6 in, 2004. (© JulianVoss-Andreae) This sculpture is a physical realization of the virtual model shown in Fig. 2,made from 100 ft square steel tubing.

the source of several important scientificinsights [29]. Sculptor Mara G. Haseltinerecently created a large-scale piece calledWaltz of the Polypeptides (2003), which por-trays the biological creation of a protein.The sculpture consists of stylized ribo-somes, with the protein represented as aribbon model [30].

I finished writing this article in thespring of 2003, and I have since createdseveral new protein sculptures. These in-clude a 5-ft-6-in-tall steel GFP (Fig. 6).This sculpture, a physical realization ofthe virtual model shown in Fig. 2, wasmade from 100 ft square steel tubing. Thenumbers that were used to enumeratethe joints of this piece can still be tracedthrough the entire length of the steel tub-ing. Certain chemical bonds (hydrogenbonds), which account for the stability ofthe molecule, are represented by ¼-inrods, connecting the square tubing in thebarrel-like structure. The spiraling move-ment of the rods runs perpendicular tothe movement of the tubing, resulting ina visual dynamic that emphasizes thebeauty of the molecule. I have also cre-ated an outdoor steel sculpture based onthe a-helix. This piece was created tohonor the memory of Linus Pauling, thediscoverer of the α-helix, and stands infront of his childhood home in Portland,Oregon. The 10-ft-tall piece, which wascommissioned by the Linus Pauling Cen-ter for Science, Peace and Health, wasmade from a 20-ft steel beam with a 12-×-12-in square cross section, cut into 15pieces. The piece was powder-coated inprimary red, complementary in color tothe green foliage embracing it. The ver-tically standing Alpha Helix appears to bebalancing on one corner, which, alongwith its location in a busy urban envi-ronment, makes it especially visually strik-ing [31].

An extensive account of how my workevolved can be found in my B.F.A. thesispaper [32], which is downloadable frommy web site [33]. The paper also containsa historical overview and an appendixwith instructions for building one’s ownprotein sculptures.

Acknowledgments

I want to thank my wife Adriana, who sparked my pas-sion for proteins when she introduced me to GFP(which she uses routinely in her neuroscience re-search) and helped me greatly in editing this manu-script. I also want to thank Jeff Baker, my fellowstudent in my first semester at the Pacific NorthwestCollege of Art in Portland, Oregon, whose denselypacked mitered woodcut sculpture reminded me of

The main page of Anton Zeilinger’s group is foundat <http://www.quantum.at>.

16. Julian Voss-Andreae, “Kohärente Moleküloptikmit Fullerenen,” Diplomarbeit (Berlin: Freie Uni-versität Berlin, 2000).

17. Lucia Hackermüller et al., “The Wave Nature ofBiomolecules and Fluorofullerenes,” Physical ReviewLetters 91 (2003) p. 90408.

18. Roger Y. Tsien, “The Green Fluorescent Protein,”Annual Review of Biochemistry 67 (1998) pp. 509–544.

19. For two examples see: <http://www.tsienlab.ucsd.edu/> and <http://www.mshri.on.ca/nagy/>.

20. See <http://www.ekac.org/>.

21. Three years after this attempt, I displayed both“sculptures” under the title Failed Biosynthesis, be-cause by then I had successfully completed a GFPsculpture in steel (Fig. 6), which I could show to-gether with the failed ones.

22. L. Skjeldal et al., “Refined Structure and MetalBinding Site of the Kalata B1 Peptide,” Archives ofBiochemistry and Biophysics 85 (2002) pp. 142–148.

23. Vivienne B. Gerritsen, “The Protein with a Topo-logical Twist,” Protein Spotlight 20 (2002); <http://us.expasy.org/spotlight/articles/sptlt020.html>.

24. Warren R. Randall et al., Manual of Oregon Treesand Shrubs (Corvallis, OR: O.S.U. Book Stores, 1990)p. 71.

25. Photos of those devices and of the cutting pro-cedure are published in my B.F.A. thesis paper, “Pro-tein Sculptures” (Portland, OR: Pacific NorthwestCollege of Art, 2004) and on my web site <http://www.JulianVossAndreae.com> (“work” → “archive”→ “Protein Project II”).

26. Because the pieces correspond to the peptideunits, their lengths should be identical in an accu-rate model of a protein.

27. Herbert Read, The Philosophy of Modern Art (Lon-don: Faber & Faber, 1952) p. 201.

28. The World Index of Molecular Visualization Resourcesweb site <http://www.molvisindex.org> gives anoverview under “Physical Molecular Models and Mo-lecular Sculpture.”

29. Information on the history of the visualizationof biological macromolecules can be found on EricMartz and Eric Francoeur’s interesting web site<http://www.umass.edu/microbio/rasmol/history.htm>.

30. See <http://calamara.com>.

31. The making of the piece was featured by the Ore-gon Public Broadcasting television series Oregon ArtBeat (6 May 2004). A clip can be downloaded frommy web site [25] (“work” → “archive” → “Protein Pro-ject V” → “click here”).

32. See ref. [25].

33. See ref. [25] (“work” → “archive” → “BFA The-sis Paper”).

34. Rief and Grubmüller [12].

Received 27 May 2003.

Julian Voss-Andreae is a sculptor based inPortland, Oregon. He did his graduateresearch in quantum physics in Anton Zei-linger’s lab in Vienna, Austria, after which heattended the Pacific Northwest College of Art,graduating with a BFA in sculpture in 2004.

a globular protein and led me to explore the rendi-tions of proteins as such sculptures. Michael May andLinda Wysong read the manuscript and provided mewith helpful comments. Helmut Grubmüller andMatthias Rief were kind enough to let me use theirinstructive picture of BPTI (Color Plate D No. 2a–d) from their article [34]. The Douglas firs used inthe α-helix sculptures were a kind gift from RogerCone.

References and Notes

1. Quoted in Herbert Read, A Concise History of Mod-ern Sculpture (New York: Praeger, 1964) p. 30.

2. D’Arcy W. Thompson, On Growth and Form, J.T.Bonner, ed., Abridged Ed. (Cambridge, U.K.: Cam-bridge Univ. Press, 1961).

3. Jack Burnham, Beyond Modern Sculpture: The Effectsof Science and Technology on the Sculpture of This Cen-tury (New York: George Braziller, 1968) p. 77.

4. H.M. Berman et al., “The Protein Data Bank,” Nu-cleic Acids Research 28 (2000) pp. 235–242 <http://www.rcsb.org/pdb/>.

5. If a physical object is referred to as being “n-di-mensional”(or “nD”), it means that it extends sig-nificantly only into n of the 3 dimensions of physicalspace. Its extent into the remaining 3-n dimensionsis negligible compared with the extent into the othern dimensions, i.e. smaller by some orders of magni-tude. Therefore, a “three-dimensional object” hascomparable length, width and height. A “two-di-mensional object” stretches out mostly into two di-mensions, and a “one-dimensional object” is muchlonger than it is wide and high.

6. William K. Purves and Gordon H. Orians, Life: TheScience of Biology (Sunderland, U.K.: Sinauer Associ-ates Inc., 1987) pp. 70–71.

7. The program performs successive multiplicationsof Euler matrices in order to rotate the coordinatesystem describing the orientation of the materialalong the protein backbone. The cutting angles arecomputed from the azimuth and the polar angle ateach point.

8. I use the Protein Data Bank [4] <http://www.rcsb.org/pdb/>. After a suitable protein has beenfound it can be viewed three-dimensionally (by using“Explore” → “View structure” → “Quick pdb”) anddownloaded as a set of Cartesian coordinates for eachatom in the molecule.

9. Carl Brändén and John Tooze, Introduction to Pro-tein Structure (New York: Garland Publishing, 1991)p. 4.

10. The Cα atoms are denoted “CA” in the .pdb filesin the Protein Data Bank.

11. The name of the Protein Data Bank file contain-ing the data is “1bpi.pdb”.

12. Matthias Rief and Helmut Grubmüller, “Kraft-spektroskopie von einzelnen Biomolekülen,”Physikalische Blätter 57, Heft 2, 55–61 (2001).

13. F. Yang, L.G. Moss and G.N. Phillips, “The Molec-ular Structure of Green Fluorescent Protein,” NatureBiotechnology 14 (1996) p. 1246.

14. Vivienne B. Gerritsen, “Mirror Mirror on the WallWho Is the Greenest of Us All?,” Protein Spotlight 11(2001) <http://us.expasy.org/spotlight/articles/sptlt011.html>.

15. Markus Arndt, Olaf Nairz, Julian Voss-Andreae,Claudia Keller, Gerbrandt van der Zouw and AntonZeilinger, “Wave-Particle Duality of C60 Molecules,”Nature 401 (1999) pp. 680–682; see <http://www.quantum.univie.ac.at/research/matterwave/c60>.

Voss-Andreae, Protein Sculptures 45

ART

AND

BIO

LOG

Y