Direct observation of native and unfolded glucose oxidase structures by scanning tunnelling microscopy

J. CHEM. SOC. FARADAY TRANS., 1994, 90(14), 2057-2060 2057

Direct Observation of Native and Unfolded Glucose Oxidase Structures by Scanning Tunnelling Microscopy

Qijin Chi, Jingdong Zhang, Shaojun Dong* and Erkang Wang* Laboratory of Electroana Iytical Chemistry , Changchun Institute of Applied Chemistry , Chinese Academy of Sciences, Changchun, Jilin 130022, P. R . China

Native and unfolded glucose oxidase (GOD) structures have been directly observed with scanning tunnelling microscopy (STM) for the first time. STM images show an opening butterfly-shaped pattern for the native GOD. W h e n GOD molecules are extended on anodized, highly ordered pyrolytic graphite (HOPG), a helical structure composed of double-stranded chains was obtained under STM. These results are in good agreement with pre- vious description of the GOD molecular structure. A simple model of the unfolding process for GOD molecules was proposed to explain these observations. Electrochemical evidence was provided to support the results obtained with STM and t h e proposed model.

In contrast to other microscopic techniques such as SEM, transmission electron microscopy (TEM) and scanning TEM (STEM), STM is capable of working in a wide variety of environments either at room temperature in air or even in aqueous solution.''2 This advantage as well as the superior resolution means that STM has been extensively used not only in imaging various solid surfaces but also in revealing the structural details of biological materials3 In the latter, a series of exciting achievements have been made and devoted to deoxyribonucleic acid and recA-DNA com- plexes.' '*' High-resolution STM images of DNA have been obtained in vacuum: in and in liquid environ- m e n t ~ , ~ , ' ~ and structural features of the double helix have been discerned,6 even down to atomic r e s~ lu t ion .~ Some reports discuss STM imaging of enzyme^,'^-^^ a typical example being phosphorylase kinase which was shown to have a bilobate structure.' 8719 In addition, virus particles, polypeptides, molecular bilayers, the purple membrane, amino acids, polysaccharides and many other proteins have been imaged with varying degrees of r e s~ lu t ion .~

Although the tunnelling mechanism through biological macromolecules is not yet fully understood, the major obs- tacle to imaging biological specimens is not tunnelling through the sample but rather fixation of the sample to the substrate.21 Several methods have been developed to over- come the fixation problem, which include covalently binding the sample to the substrate6 using biological materials to form aggregates22 and electrochemical deposition.' 0,20i23

However, unlike the special properties of the biological speci- mens exploited in those studies, it is demonstrated here that GOD can be successfully imaged with STM when the enzyme is deposited on freshly cleaved HOPG or irreversibly adsorbed onto an anodized HOPG surface.

GOD is a flavin enzyme with flavin adenine dinucleotide (FAD) as the redox prosthetic group, its biological function being to catalyse glucose to form gluconolactone. From bio- chemical s t ~ d i e s ~ ~ * ~ ~ GOD is known to be a structurally rigid glycoprotein of ca. 160000 Da t and consists of two identical polypeptide chains, each containing a FAD redox centre. However, direct observation of its structural features has never been performed.

Experimental GOD (from Aspergillus, EC 1.1.3.4) was obtained from Sigma (Type 11, 35 300 U g- ', G-6125). Other reagents used were of

t 1 Da = 1 u.

analytical-reagent grade. All solutions were prepared with doubly distilled water. GOD was dissolved in doubly distilled water and the exact concentration of the enzyme solution was determined by spectrophotometric assays at 452 nm using a molar absorptivity of 21.6 dm3 mo1-' cm-1.26 Electrochemical experiments were carried out on an FDH 3204 potentiostat (Shanghai) with a Gould Series-60000 X-Y recorder (Shenyang, China). A three-electrode system was employed with an Ag/AgCl (saturated KC1) reference elec- trode, a platinum plate auxiliary electrode and an HOPG working electrode. All buffer solutions were purged with nitrogen gas before electrochemical experiments. A Varian DMS-90 UV-VIS spectrophotometer was used for spectro- photometric assays. STM imaging was performed under a normal atmosphere at room temperature (20 & 2°C) with a TO~OMETRIX TMX 2000; this STM instrument was com- bined with a monitor which was used in observing the dis- tance between the sample and the tip. The air humidity was kept at 40-60%. The Pt/Ir (80/20) tips were mechanically formed by cutting. Scanning was in the constant-current mode except for the images of bare anodized HOPG, at a bias voltage of 0.2-0.4 V and a currrent of 1.0-1.5 nA. Post- image processing of all images involved low-pass filtering to remove high-frequency noise without two-dimensional Fourier transformation. GOD was deposited in a 15 mm3 droplet of 4.3 x mol dm3 enzyme solution on the freshly cleaved HOPG surface. The sample was then dried at room temperature (drying time 40-60 min) and finally washed carefully with doubly distilled water. Another sample was prepared by the same process except that the freshly cleaved HOPG was firstly anodized at +2.0 V for 2 min.

Results and Discussion Fig. 1 shows the absorption spectrum of 4.3 x mol dm3 GOD solution. Absorption is significant in the visible region with maximum values at 380 and 452 nm, but above 452 nm the absorption decreases sharply. This is consistent with a previous report.26 When p-hydroquinone was used as a medi- ator, electrochemical experiments demonstrated that GOD can rapidly catalyse glucose to gluconolactone. Therefore, the present results confirm that the GOD used retains its native structure with biological activity.

GOD can be adsorbed on the freshly cleaved HOPG to form a monolayer ; usually molecular images are easily obtained at the edge of the HOPG surface, and possibly the fixation of GOD molecules is more stable in these regions. A typical image of G O D molecules is shown in Plate l(a). Four

Dow

nloa

ded

by D

TU

Lib

rary

on

17 J

uly

2012

Publ

ishe

d on

01

Janu

ary

1994

on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/F

T99

4900

2057

View Online / Journal Homepage / Table of Contents for this issue

2058 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90

0.09

0.08

0.07

0.06

a 2 0.05 5 2 4 0.04

0.03 1 0.02 '

0.01 '

0.00 '

300 400 500 6 0 0 A/nm

Fig. 1 tion

Absorption spectrum of 4.3 x lop6 mol dm-3 GOD solu-

individual enzyme molecules are clearly discerned and are distributed uniformly on the substrate surface. Many scans were performed in other regions of the substrate and this pattern was found to be reproducible. The scanning range was then limited to 31 x 31 nm2 or smaller, which allows detailed observation of an enzyme molecule [Plate l(b)]. Plate l(c) was obtained from Plate l(b) by rotating the sample by 90" and reducing the scan area; the corresponding three-dimensional image is shown in Plate l(d). As visualized from Plate l(c) and (4, the individual enzyme molecule exhibits a butterfly-shaped structure containing two sym- metric wings and a depressed bridge or centre. This pattern is similar to the STM image of phosphorylase kinase,I8 but the former is smaller. Quantitative size determinations were made with the TOPOMETRIX Instrument software. Fig. 2 shows typical line profiles obtained through the various direction noted in Plate l(c). The largest length across the wing tips [see Fig. 2(4, (b)] is 13.8 f 0.5 nm ( n = 12) and the largest width across the bridge [Fig. 2(c)] is 18.1 f 0.6 nm ( n = 12). The average length of a wing (parallel to the edge) is 12.2 0.4 nm ( n = 20) and the average width of each wing [perpendicular to the edge; Fig. 2(4] is 8.9 f 0.4 ( n = 20). The width of the bridge was measured (parallel to the edge through the bridge) as 4.7 f 0.3 nm ( n = 16). The average height of the molecule was measured as 1.27 & 0.06 nm ( n = 15) on the basis of the calculated average displacement of the probe tip from the substrate while the molecule was scanned. As different electronic work functions exist between the substrate (HOPG) and the adsorbed enzyme, the mea- sured vertical distance will not be consistent with the physical height. In other words, the absolute thickness of the molecule cannot be accurately obtained from these STM measure- ments. However, the relative height shown here is expected to be proportional to the absolute thickness. Thus, the size of a GOD molecule in this study can be approximately described as 18 x 12 x 1.3 nm3, but the exact dimensions of an individ- ual molecule are still unknown, although it is considered in general to be ca. 9.0 nm in diameter.24 In addition, the line

L rn

2.95 w II 0.93 ' I Y

0.00 11.92 23.84

0.64' I I 0.00 11.50 22.00

distance/nm (a), (b), (c) and (dj sectional cuts obtained from Plate l(c) Fig. 2

along the direction of ab, cd, ef and gh, respectively.

profile shown in Fig. 2(c) indicates clearly that the bridge portion of the molecule between two wings appears to be depressed; this result further supports the above description of a butterfly-shaped pattern.

In previous studies, HOPG was activated by various means, such as chemical ~xida t ion ,~ ' laser activation,28 elec- trochemical ~ x i d a t i o n ~ ~ , ~ ' and reaction with dioxygen at ele- vated t e m p e r a t ~ r e , ~ ' . ~ ~ to create active sites and to accelerate heterogeneous electron transfer between HOPG and the redox systems. Here, electrochemical pretreatment at a high positive voltage was used to produce active sites on the HOPG surface. Plate 2 shows the STM images of bare HOPG after anodizing it at +2.0 V for 2 min; clearly nanometre-sized pits were formed in the substrate plane [Plate 2(b)]. These etching pits provide effective active sites for the adsorption of GOD on the HOPG. The GOD mol- ecules were adsorbed irreversibly onto the anodized HOPG to yield a stable sample. A series of STM images with differ- ent scan ranges are shown in Plate 3. These images are obvi- ously different from those shown in Plate 1, and indicate that the tertiary structure of the enzyme has been lost on this surface. As shown in Plate 3(a)-(c), several structures, includ- ing chains, clusters and fragments, exist simultaneously on this surface. In addition, sometimes a tetramer of unfolded GOD molecules could also be observed, as shown in the left- hand side of Plate 3(d). In these various structures, a typical example is the isolated fibre shown in Plate 3(e) and (f), which can be clearly discerned to be composed of two poly- peptide chains twisted into a rope-like structure. A cross-

Dow

nloa

ded

by D

TU

Lib

rary

on

17 J

uly

2012

Publ

ishe

d on

01

Janu

ary

1994

on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/F

T99

4900

2057

View Online

J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90

Plate 1 STM images of GOD adsorbed on the freshly cleaved HOPG obtained with different scan scopes. (a) and (b) topographic views; (c) obtained from (b) after rotating it by 90" and reducing the scan area; (d) three-dimensional image of (c). U = 0.3 V; I = 1.2 nA; scan area = 88.33 x 88.33 nm2 (a), 31.42 x 31.42 nm2 (b), 22.82 x 22.82 nm2 (c), (d).

Plate 2 Topographic views of bare HOPG electrochemically pretreated at +2.0 V for 2 min obtained in constant-height mode. U = 0.4 V; scan area = 120 x 120 nm2 (a), 10 x 10 nm2 (b).

Q, Chi et al. (Facing p. 2058)

Dow

nloa

ded

by D

TU

Lib

rary

on

17 J

uly

2012

Publ

ishe

d on

01

Janu

ary

1994

on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/F

T99

4900

2057

View Online

J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90

Plate 3 scan area = lo00 x loo0 nm2 (a), 400 x 400 nm2 (b), 200.08 x 200.08 nm2 (c), 100 x 100 nm2 (d) , 70.17 x 70.17 nm2 (e) and 45 x 45 nm2 (f).

STM images of GOD adsorbed on the anodized HOPG at various scan scopes. ( a ) - ( f ) Topographic views. U = 0.2 V; I = 1.0 nA;

Dow

nloa

ded

by D

TU

Lib

rary

on

17 J

uly

2012

Publ

ishe

d on

01

Janu

ary

1994

on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/F

T99

4900

2057

View Online

J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 2059

5*50* 2.02 3.76

0.00 22.98 45.96

distance/nm

Fig. 3 (a), (b) and (c) line profiles obtained from Plate 3 (d) and (f) along the directions of ab, cd and ef, respectively

section of Plate 3(e) along the ab direction through the chain is shown in Fig. 3(a), and the helical pitch was measured as 15.1 f 0.7 nm ( n = 15). The chain width was determined by line scans perpendicular to the chain [e.g. Fig. 3(b)]; the average value is 12.5 f 1.4 nm ( n = 22). Size determination of each strand chain was made by drawing a section through ef shown in Plate 3(f); the result is given in Fig. 3(c). The diam- eters of two strand chains are 8.0 and 6.3 nm, respectively. After deducting the space influence, these two strand chains can be considered to be the same size. Therefore, the rope- like pattern observed is composed of two identical chains and represents the unfolded form of an individual enzyme mol- ecule.

In order to understand these observations better, a simple model is proposed in Fig. 4. Fig. 4(a) shows the native struc- ture of the enzyme, which has tertiary and quaternary struc- tures with a folded form. When the enzyme is adsorbed onto

G. GL

G . J / I \

e HOPG e HOPG e HOPG e HOPG

(b 1 ( c ) ( d 1 (e 1 Fig. 4 Proposed model of the extending process of the GOD mol- ecules on the anodized HOPG. G, GL and large solid spots represent glucose, gluconolactone and FAD, respectively.

the freshly cleaved HOPG, the enzyme molecules retain their native or approaching native structure as shown in Fig. 4(a) and (b) owing to the weak enzymesurface interaction. There- fore, under this situation a butterfly-shaped pattern com- posed of two wings and a depressed bridge could be observed from STM imaging. Each wing represents the folded form of a polypeptide chain, whereas the bridge denotes the non- covalent bonding part of two polypeptide chains of a GOD molecule. In contrast, a strong interaction between enzyme and the substrate surface occurs when GOD is adsorbed onto the anodized HOPG surface, which results in GOD mol- ecules gradually unfolding, as shown in Fig. 4(b)-(e). The extent of unfolding of the enzyme molecules depends on the degree of enzyme-surface interaction. Since the anodized HOPG surface is not uniformly flat (see Plate 2), unlike the untreated HOPG surface, the strength of the enzyme-surface interaction must vary. Therefore, when the G O D molecules are adsorbed onto the anodized HOPG surface, at least four kinds of structures are expected to be present: (1) the poly- peptide chain of GOD fully extended [Fig. 4(e)], giving an individual chain structure ; (2) when the fully extended poly- peptides aggregate together, trimers or tetramers of the unfolded molecules would be formed; (3) the native GOD molecules extend only partly as shown in Fig. 4(c) and (d) or the unfolded molecules coil randomly, giving an intermediate or shorter chain structure; (4) a strong interaction between the enzyme molecules and the substrate may also result in the enzyme molecules being broken to form fragments. Indeed, all these structures were observed with STM, as shown in Plate 3. Clearly the typical structure shown in Plate 3(e) and ( f ) originate from the fully unfolded form of an individual GOD molecule. On the other hand, it is well known that the redox centre (FAD) of the GOD molecule is embedded deep in the enzyme in the native structure. The distance between either of its two FAD centres and the electrode surface exceeds the distance across which electrons are transferred at measurable rates, hence direct electron transfer between the native enzyme and the electrode surface cannot occur. A large number of attempts have been made to achieve direct electron exchange between GOD and various elec- t r ~ d e s ; ~ ~ - ~ ' however, when the GOD molecules are adsorbed on the HOPG surface and gradually extend, the redox centre of the enzyme is also gradually exposed to the HOPG surface, the distance between the redox centre of enzyme and the substrate surface becoming shorter and shorter [see Fig. 4(b)-(e)] ; thus direct electron transfer from adsorbed GOD to the HOPG surface is expected to be pos- sible. In this study, electrochemical experiments confirm this prediction. When G O D is deposited on the untreated HOPG, no electrochemical signals resulting from direct elec- tron transfer can be measured, indicating that the adsorbed GOD retains its native or approximately native structure and the distance between the redox centre and the HOPG surface remains larger than the critical distance under this situation. In contrast, the well defined cyclic voltammograms of GOD adsorbed on anodized HOPG were obtained and shown in Fig. 5. A couple of redox peaks appear at -0.4 to -0.5 V us. Ag/AgCl, which result from the direct reduction and oxida- tion of the adsorbed GOD and can be described by the fol- lowing equation:

GOD(FAD) + 2e + 2H+ e GOD(FADH,) (1)

As expected, the GOD adsorbed on the anodized HOPG surface could not catalyse its substrate (glucose) oxidation, indicating that the adsorbed GOD had been denatured to lose biological activity owing to extending to an unfolded form. Consequently the electrochemical results shown here lend further support to the STM observations described

Dow

nloa

ded

by D

TU

Lib

rary

on

17 J

uly

2012

Publ

ishe

d on

01

Janu

ary

1994

on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/F

T99

4900

2057

View Online

2060 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90

1 " ' 1 ' 1 ' ' . 1

0.3 0.1 -0.1 -0.3 4 . 5 4 . 7 E/V vs. AgjAgCl

Fig. 5 Cyclic voltammograms of bare anodized HOPG (dashed line) and GOD adsorbed on the anodized HOPG (solid lines) obtained in 0.1 mol dm-3 phosphate buffer (KH,PO, + Na,HPO,) (pH 6.7) with scan rates (mV s - l ) of (a) and (b) 30, (c) 50, (d) 70, (e) 100 and (f) 150

above and to the model shown in Fig. 4. Finally, it should be pointed out that the interaction between GOD and the anod- ized HOPG surface possibly includes physical and chemical interactions, but the real mechanism of interaction is not clear at present.

Conclusions (1) Glucose oxidase can be adsorbed on freshly cleaved HOPG and retains its native structure, a butterfly-shaped pattern, consisting of two symmetrical wings and a depressed bridge, which has been clearly observed by STM.

(2) In order to observe the double-stranded chains of the GOD molecule, it is necessary to make the enzyme extend fully to the unfolded form. A simple method, by adsorption of the enzyme onto the anodized HOPG surface, has been developed to prepare an unfolded GOD sample. A typical rope-like structure composed of two twisted polypeptide chains has been successfully visualized by STM.

(3) Since the redox centre of the enzyme can be exposed sufficiently to the anodized HOPG surface, direct electron transfer between the adsorbed GOD and the treated sub- strate has been achieved.

Thanks are due to Prof. Zemu Yu, Bailin Zhang and Jin Li for their technical help and useful discussions. The support of the National Natural Science Foundation of China is greatly appreciated.

References 1 G. Binnig and H. Rohrer, ZBM J. Res. Deo., 1986,30,355. 2 P. K. Hansma and J. Tersoff, J. Appl. Phys., 1987,61, R1. 3 S. R. Snyder and H. S. White, Anal. Chem., 1992,64,116R. 4 R. J. Driscoll, M. G. Youngquist and J. D. Baldeschwieler,

Nature (London), 1990,346,294.

5 T. P. Beebe Jr., T. E. Wilson, D. F. Ogletree, J. E. Katz, R. Balhorn, M. B. Salmeron and W. J. Siekhaus, Science, 1989,243, 370.

6 A. Cricenti, S. Selci, A. C. Felici, R. Generosi, E. Gori, W. Djac- zenko and G. Chirotti, Science, 1989,245, 1226.

7 D. D. Dunlap and C. Bustamante, Nature (London), 1989, 342, 204.

8 G. Lee, P. G. Arscott, V. A. Bloomfield and D. F. Evans, Science, 1989,244,475.

9 P. G. Arscott, G. Lee, V. A. Bloomfield and D. F. Evans, Nature (London), 1989,339,484.

10 S . M. Lindsay, T. Thundat, L. Nagahara, U. Knipping and R. L. Rill, Science, 1989, 244, 1063.

11 M. Amrein, A. Stasiak, H. Gross, E. Stoll and G. Travaglini, Science, 1988,240, 514.

12 M. Amrein, R. Durr, A. Stasiak, H. Gross and G. Travaglini, Science, 1989, 243, 1708.

13 R. Czajika, C. G. J. Koopal, M. C. Feiters, J. W. Gerritsen, R. J. M. Nolte and H. van Kempen, Bioelectrochem. Bioenerg., 1992, 29,47.

14 P. K. Hansma, V. B. Elings, 0. Marti and C. E. Bracke, Science, 1988,242,209.

15 S . A. Darst, E. W. Kubalek and R. D. Kornberg, Nature (London), 1989,340,730.

16 M. H. Jericho, B. L. Blackford, D. C. Dahn, C. Frame and D. Macleean, J. Vac. Sci. Technol. A, 1990,8,661.

17 L. Haggerty, B. A. Waston, M. A. Barteau and A. M. Lenhoff, J. Vac. Sci. Technol. By 1990,9, 1219.

18 R. D. Edstrom, M. H. Meinke, X. Yang, R. Yang and D. F. Evans, Biochemistry, 1989,28,4939.

19 V. B. Elings, R. D. Edstrom, M. H. Meinke, X. Yang, R. Yang and D. F. Evans, J. Vac. Sci. Technol. A , 1991,8,652.

20 R. W. Keller, D. Bear and C. Bustamante, J. Vac. Sci. Technol. B, 1991,9, 1291.

21 M. Salmeron, T. P. Beebe Jr., J. Odriozola, T. Wilson, D. F. Ogletree and W. Siekhaus, J. Vac. Sci. Technol. A , 1990,8, 635.

22 M. J. Miles, T. McMaster, J. J. Carr, A. S. Tatham, P. R. Shewry, J. M. Field, P. S. Belton, D. Jeenes, B. Hanley, M. Whittam, P. Cairns, V. I. Morris and N. Lambert, J. Vac. Sci. Technol. A, 1990,8,698.

23 S . M. Lindsay and B. Barris, J. Vac. Sci. Technol A, 1988,6,544. 24 S . Nakamura, S. Hayashi and K. Koga, Biochim. Biophys. Acta,

1976,445,294. 25 The Worthington Manual, Worthington BiochemicaI, Freehold,

26 S. Nakamura and S . Fujiki, J . Biochem., 1968,63, 51. 27 L. Porte, D. Richard and P. Gallezot, J . Microsc., 1988, 152, 515. 28 R. S. Robinson, K. Sternitzke, M. T. McDermott and R. L.

McCreery, J. Electrochem. SOC., 1991,138,2412. 29 A. A. Gewirth and A. J. Bard, J. Phys. Chem., 1988,92,5563. 30 C. A. Goss, J. C. Brumfield, E. A. Irene and R. W. Murray, Anal.

Chem., 1993,651378. 31 H. Chang and A. J. Bard, J. Am. Chem. SOC., 1990,112,4598. 32 H . Chang and A. J. Bard, J. Am. Chem. SOC., 1991,113,5588. 33 Y. Degani and A. Heller, J. Phys. Chem., 1987,91,1285. 34 P. N. Bartlett, R. G. Whitaker, M. J. Green and J. Frew, J.

Chem. Soc., Chem. Commun., 1987, 1603. 35 Y. Degani and A. Heller, J. Am. Chem. SOC., 1988,110,2615. 36 W. Schuhmann, T. J. Ohara, H-L. Schmidt and A. Heller, J. Am.

Chem. SOC., 1991,113, 1394. 37 Y. Kajiya and H. Yoneyama, J. Electroanal. Chem., 1992, 328,

259. 38 Y. Kajiya and H. Yoneyama, J. Electroanal. Chem., 1992, 341,

85. 39 C. G. J. Koopal, M. C. Feiters, R. J. M. Nolte, B. de Ruiter,

R. B. M. Schasfoort, R. Czajka and H. Van Kempen, Synth. Met., 1992,51, 397.

40 C. G. J. Koopal, M. C. Feiters, R. J. M. Nolte, B. de Ruiter and R. B. M. Schasfoort, Biosensors Bioelectronics, 1992,7,461.

NJ, 1977, pp. 37-39.

Paper 41006421; Received 1st February, 1994

Dow

nloa

ded

by D

TU

Lib

rary

on

17 J

uly

2012

Publ

ishe

d on

01

Janu

ary

1994

on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/F

T99

4900

2057

View Online