Expressing the Geobacter metallireducens PilA in Geobacter sulfurreducens Yields Pili with Exceptional Conductivity

ABSTRACT The electrically conductive pili (e-pili) of Geobacter sulfurreducens serve as a model for a novel strategy for long-range extracellular electron transfer. e-pili are also a new class of bioelectronic materials. However, the only other Geobacter pili previously studied, which were from G. uraniireducens, were poorly conductive. In order to obtain more information on the range of pili conductivities in Geobacter species, the pili of G. metallireducens were investigated. Heterologously expressing the PilA gene of G. metallireducens in G. sulfurreducens yielded a G. sulfurreducens strain, designated strain MP, that produced abundant pili. Strain MP exhibited phenotypes consistent with the presence of e-pili, such as high rates of Fe(III) oxide reduction and high current densities on graphite anodes. Individual pili prepared at physiologically relevant pH 7 had conductivities of 277 ± 18.9 S/cm (mean ± standard deviation), which is 5,000-fold higher than the conductivity of G. sulfurreducens pili at pH 7 and nearly 1 million-fold higher than the conductivity of G. uraniireducens pili at the same pH. A potential explanation for the higher conductivity of the G. metallireducens pili is their greater density of aromatic amino acids, which are known to be important components in electron transport along the length of the pilus. The G. metallireducens pili represent the most highly conductive pili found to date and suggest strategies for designing synthetic pili with even higher conductivities.

IMPORTANCE e-pili are a remarkable electrically conductive material that can be sustainably produced without harsh chemical processes from renewable feedstocks and that contain no toxic components in the final product. Thus, e-pili offer an unprecedented potential for developing novel materials, electronic devices, and sensors for diverse applications with a new "green" technology. Increasing e-pili conductivity will even further expand their potential applications. A proven strategy is to design synthetic e-pili that contain tryptophan, an aromatic amino acid not found in previously studied e-pili. The studies reported here demonstrate that a productive alternative approach is to search more broadly in the microbial world. Surprisingly, even though G. metallireducens and G. sulfurreducens are closely related, the conductivities of their e-pili differ by more than 3 orders of magnitude. The ability to produce e-pili with high conductivity without generating a genetically modified product enhances the attractiveness of this novel electronic material.
As recently reviewed (11)(12)(13)(14), various theoretical modeling approaches have suggested different mechanisms to account for this unique long-range conductivity in a biological protein. Final resolution of the actual mechanism is likely to require experimental determination of the e-pilus structure. However, resolution of the G. sulfurreducens e-pilus structure will be technically challenging for such a thin (3-nm) filamentous structure.
A major impetus for developing a better understanding of the mechanisms for e-pili conductivity is the possibility that this will lead to strategies for developing synthetic e-pili with enhanced functions, such as higher conductivity. For example, experimental evidence has clearly demonstrated the important role of aromatic amino acids in promoting e-pili conductivity. X-ray diffraction demonstratedstacking of aromatic amino acids, which has been proposed to confer metallic-like conductivity along the length of e-pili (1,13). Altering the degree ofstacking by changing pH (1,3) or genetic manipulation (2) leads to changes in e-pili conductivity directly related to the degree ofstacking (13). Eliminating thestacking of aromatic amino acids with genetic manipulation eliminates the charge propagation along the pili, which can be documented with electrostatic force microscopy (15). From these experimental results, it was possible to devise strategies to either tune down the conductivity of G. sulfurreducens e-pili by removing aromatic amino acids from the e-pilus monomer PilA (2,16) or to substantially increase the conductivity by adding tryptophan (5).
An alternative approach to genetic manipulation of PilA for producing e-pili with different conductivities is to examine the conductivity of pili of other microorganisms. The pili of Geobacter uraniireducens are more than 100-fold less conductive than G. sulfurreducens pili (17). This was attributed to the much longer length of the PilA monomer of G. uraniireducens, which may prevent aromatic amino acids from packing sufficiently tight for effective electron transport. A similar explanation has been suggested for the poor conductivity of Pseudomonas aeruginosa pili (18).
The poor conductivity of G. uraniireducens pili is associated with phenotypes for extracellular electron transfer that are markedly different than those for G. sulfurreducens. Whereas G. sulfurreducens requires direct contact to reduce Fe(III) oxides (19), G. uraniireducens produced an electron shuttle for long-range electron transport (17). Furthermore, G. uraniireducens was not capable of producing high current densities (20). G. sulfurreducens requires conductive e-pili to produce high current densities (9,10,21). These results suggested that G. uraniireducens does not utilize e-pili for long-range electron transport. However, PilA sequence analysis has suggested that other Geobacter species and closely related microorganisms may have pili that are electrically conductive (22).
It has been indirectly inferred that the pili of Geobacter metallireducens are electrically conductive, as a mutant strain in which the gene for PilA was deleted was ineffective in extracellular electron transfer to Fe(III) oxides (23,24) or other cells (8,25,26). G. metallireducens was the first Geobacter species isolated (27) and is one of the most effective Fe(III) oxide-reducing Geobacter species (28). It can produce high current densities (20), and it has the ability to forge direct electrical connections with methanogenic microorganisms (25,26) for direct interspecies electron transfer (DIET). Specific expression of pili during growth of G. metallireducens on Fe(III) oxides, but not on chelated Fe(III), was the first indication that pili might be important in Fe(III) oxide reduction in Geobacter species (29). Here, we report that the pili of G. metallireducens are much more conductive than the Geobacter pili that have been previously examined and even more conductive than currently available synthetically designed pili.

RESULTS AND DISCUSSION
A G. sulfurreducens strain that produces G. metallireducens pili. Heterologous expression of pili from other organisms in G. sulfurreducens has been shown to facilitate rapid screening of the conductivity of diverse pili via evaluation of current densities produced on an anode in a common host, and to provide an abundant source of pili for additional analysis (2,17,18). Therefore, the PilA gene of G. sulfurreducens was replaced with the PilA gene of G. metallireducens via methods previously developed for heterologous expression of other PilA sequences (2,17,18). The G. metallireducens PilA contains two fewer amino acids than the PilA of G. sulfurreducens and has a higher content of aromatic amino acids, with a tyrosine at position 50, a histidine at position 54, and a phenylalanine at position 56, positions where there are nonaromatic amino acids in the G. sulfurreducens PilA (Fig. 1A). In addition, there is a phenylalanine in the G. metallireducens PilA at position 32, whereas there is a tyrosine at this position in the G. sulfurreducens PilA.
G. sulfurreducens requires conductive pili in order to produce high current densities. Strain MP produced current densities ( Fig. 2A) comparable to the previously reported (2) current densities generated under similar conditions for a control strain expressing the G. sulfurreducens wild-type PilA gene. The high current densities were associated  Highly Conductive Pili from Geobacter metallireducens ® with thick biofilms, with a staining response suggesting that the cells were viable throughout (Fig. 2B). This result is consistent with the need for metabolically active cells at a distance from the anode to contribute electrons to the anode in order to produce high current densities, and this is only possible when there is long-range electron transport through the biofilms.
Pilus conductivity. Individual pili bridging the nonconductive gap between electrodes on an electrode array (Fig. 4A) had a height of 3 nm (Fig. 4B). This diameter is comparable to that of the e-pili of G. sulfurreducens (3). Conductivity was determined at pH 7 for physiological relevance. The individual pili had a linear ohmic response to current over a small, physiologically relevant voltage span (Fig. 4C). The relatively high currents (nanoamperes) through the pili necessitated that conductivity be evaluated over a lower voltage range than in previous studies (3) to avoid damaging the sample. The conductivity of the pili was 277 Ϯ 18.9 S/cm (mean Ϯ standard deviation for three pili).
The conductivity of the G. metallireducdens pili at pH 7 was 5,000-fold higher than the conductivity of G. sulfurreducens pili prepared under similar conditions and nearly 1 million-fold higher than the conductivity of G. uraniireducens pili (Fig. 5A). It was also higher than the previously reported (5) conductivity of a synthetic pilus in which a phenylalanine and a tyrosine in the G. sulfurreducens PilA were replaced with tryptophan (Fig. 5A). Chemical treatments of pili, such as preparing them at pH 2, can dramatically increase their conductivity (3), but even at pH 7 the pili of G. metallireducens are more conductive than the G. sulfurreducens pili prepared at pH 2 (0.2 S/cm) and nearly as high as the conductivity of synthetic pili (391 S/cm) at pH 2. Estimates of conductivity along the length of G. sulfurreducens pili prepared in solvents and dried (4) were somewhat higher (2.85 S/cm) than the estimate for G. sulfurreducens obtained without such harsh chemical conditions (Fig. 5A) but were still orders of magnitude lower than the conductivity of the G. metallireducens pili. Rhodopseudomonas palustris filaments of unknown composition that are thought to be involved in extracellular electron transfer (30) had much lower conductivities (0.053 S/cm), but these filaments were also chemically fixed and dried, which may have affected their conductivity.
Implications. These results demonstrated a remarkable nearly million-fold range in the conductivity of pili within Geobacter species and suggest that the density of aromatic rings in the pilus structure is a key factor in determining pilus conductivity. These findings provide a basis for predicting the conductivity of other pili and for the design of synthetic pili with even higher conductivity. The high conductivity of the G. metallireducens e-pili suggests that they may be an attractive material for the construction of conductive materials, electronic devices, and sensors that may be developed with e-pili.
Previous studies have demonstrated that aromatic amino acids contribute to the conductivity of G. sulfurreducens pili (1)(2)(3)(4)(5)13). The substantially higher conductivity of the G. metallireducens pili is consistent with this concept. There is a clear relationship between the conductivity along the length of individual pili estimated with the same method and the density of aromatic rings in the pili (Fig. 5B). The additional aromatic amino acids may provide more or better paths for electron transport. The conductivityaromaticity relationship suggests that the design of synthetic e-pili with even more aromatic amino acids could yield an even better conducting material. It is also possible that nature has already produced e-pili that are more aromatic amino acid dense than G. metallireducens pili. Thus, further prospecting for e-pili in the microbial world is warranted.
The physiological advantage, if any, to G. metallireducens of expressing e-pili that are more conductive than those of G. sulfurreducens is not yet clear. It has been estimated that G. sulfurreducens pili are sufficiently conductive that just two e-pili could accommodate maximum rates of extracellular electron transfer to Fe(III) oxides, and cells typically produce more than 20 e-pili (17). Therefore, it is not surprising that expression of G. metallireducens e-pili in G. sulfurreducens did not yield a strain that reduced Fe(III) Highly Conductive Pili from Geobacter metallireducens ® oxide faster than the control strain constructed in the same manner but expressing the wild-type G. sulfurreducens pili. However, there is a minimum e-pili conductivity that is required for long-range electron transport to Fe(III) oxides, as evidenced by the fact that G. sulfurreducens did not effectively reduce Fe(III) oxides when expressing poorly conductive pili from G. urannireducens (17), Pseudomonas aeruginosa (18), or synthetically designed pili (2).
It also appears that increasing the conductivity of e-pili beyond some minimum threshold does not increase the capacity for current production. Neither G. metallireducens (20) nor the G. sulfurreducens MP strain (this study) produced higher current densities than G. sulfurreducens expressing the wild-type PilA. However, some e-pili conductivity is required for high current densities, as evidenced by the fact that G. sulfurreducens expressing poorly conductive pili (2,17,18) or no pili (9, 31) produces low current densities.
It is possible that higher pili conductivities could be beneficial for DIET. For example, a more conductive electrical connection between electron-donating Geobacter species and electron-accepting methanogens (25,26) could facilitate the delivery of electrons at potentials low enough to support methanogenesis. Detailed studies on the role of pili conductivity in DIET are under way.