high arsenic content perhaps making the success of this risky venture more likely.
Aerobic enrichment cultures were setup using a mineral salts medium, glucose and
arsenate in lieu of phosphate. These were taken through several transfers and
decimal dilutions and plated on agar from which a colony was isolated. A growth
curve was followed by inoculating the isolate, strain GFAJ-1, into liquid medium
lacking either arsenate or phosphate. No growth was noted in these controls, but
addition of 40 mM arsenate or 1.5 mM phosphate resulted in obvious enhanced
growth. It was clear from a number of experimental protocols that arsenic had
entered the cells, especially in the arsenate replete condition (+As/-P). NanoSIMS
analysis of extracted DNA indicated it was denser and had a higher As/P ratio (13.4)
than that of the –As/+P condition (6.9). Finally, X-ray spectroscopy (EXAFS) of
DNA in whole cells (+As/-P condition) as modeled on As substituting for P closely
aligned with the observed experimental spectra. Moreover, these data further indicated that the arsenic atom was bonded to oxygen (i.e., arsenate), which in turn was
bonded to the carbon of the deoxyribose sugar. This ruled out the possibility of direct
arsenic to carbon bonds as occurs in such molecules as arseno-betaine. The paper
underwent internal review and then was submitted to Science. It was favorably
reviewed and published online on December 2, 2010.
For reasons I will not delve into, high praise soon turned into severe criticism.
The biggest viable criticism we faced was that arsenate-carbon ester bonds are
inherently unstable and break apart quickly in aqueous medium.
Laboratories from around the world requested strain GFAJ-1 for testing. We
plated out the cultures and shipped them to certain bona fide laboratories, but instead
of using arsenate, for safety reasons we replaced with phosphate. I was aware of the
possibility that removal of the arsenate selective factor could work against us. We
sent the microbe to Rosie Redfield at the University of Vancouver, as well as Julia
Vorholt, at the ETH in Zurich. Alas, neither of these labs detected appreciable
arsenic in the DNA they extracted from grown cultures of GFAJ-1 (Reaves et al.
2012; Erb et al. 2012), although Elias et al. (2012) reported the bacterium’s interesting ability to discriminate in favor of phosphate uptake against a high background
of arsenic, while Wu et al. (2018) reported on the exceptional arsenic-resistance
capacity of strain GFAJ-1. Scavenging of phosphate from ribosome breakdown was
suggested as the source of GFAJ’s phosphate as evidenced by experimental work
with E. coli (Basturea et al. 2012).
Our refutation of these papers is that the efforts by Reaves et al. and Erb et al.
employed additional treatments to the extracted DNA before analyses. After an
initial phenol chloroform extraction, Reaves et al. purified the DNA with RNase A
and Proteinase K followed by phenol chloroform extraction and ethanol precipitation. Erb et al. did not use RNase A but repeated 70% ethanol rinses. The DNA was
then resuspended in water before it was analyzed. Hence any susceptible arsenateesters would have been broken down by the extraction process itself. We did not
employ such methods, relying on NanoSIMS and X-ray spectroscopy for our results.
Indeed, in our paper we made note of the instability of the molecule in water and
suggested that it could have been stabilized internally by such molecules as
poly-β-hydroxybutyrate. Moreover, Wolfe-Simon pointed out that DNA within
680
R. S. Oremland
Aerobic enrichment cultures were setup using a mineral salts medium, glucose and
arsenate in lieu of phosphate. These were taken through several transfers and
decimal dilutions and plated on agar from which a colony was isolated. A growth
curve was followed by inoculating the isolate, strain GFAJ-1, into liquid medium
lacking either arsenate or phosphate. No growth was noted in these controls, but
addition of 40 mM arsenate or 1.5 mM phosphate resulted in obvious enhanced
growth. It was clear from a number of experimental protocols that arsenic had
entered the cells, especially in the arsenate replete condition (+As/-P). NanoSIMS
analysis of extracted DNA indicated it was denser and had a higher As/P ratio (13.4)
than that of the –As/+P condition (6.9). Finally, X-ray spectroscopy (EXAFS) of
DNA in whole cells (+As/-P condition) as modeled on As substituting for P closely
aligned with the observed experimental spectra. Moreover, these data further indicated that the arsenic atom was bonded to oxygen (i.e., arsenate), which in turn was
bonded to the carbon of the deoxyribose sugar. This ruled out the possibility of direct
arsenic to carbon bonds as occurs in such molecules as arseno-betaine. The paper
underwent internal review and then was submitted to Science. It was favorably
reviewed and published online on December 2, 2010.
For reasons I will not delve into, high praise soon turned into severe criticism.
The biggest viable criticism we faced was that arsenate-carbon ester bonds are
inherently unstable and break apart quickly in aqueous medium.
Laboratories from around the world requested strain GFAJ-1 for testing. We
plated out the cultures and shipped them to certain bona fide laboratories, but instead
of using arsenate, for safety reasons we replaced with phosphate. I was aware of the
possibility that removal of the arsenate selective factor could work against us. We
sent the microbe to Rosie Redfield at the University of Vancouver, as well as Julia
Vorholt, at the ETH in Zurich. Alas, neither of these labs detected appreciable
arsenic in the DNA they extracted from grown cultures of GFAJ-1 (Reaves et al.
2012; Erb et al. 2012), although Elias et al. (2012) reported the bacterium’s interesting ability to discriminate in favor of phosphate uptake against a high background
of arsenic, while Wu et al. (2018) reported on the exceptional arsenic-resistance
capacity of strain GFAJ-1. Scavenging of phosphate from ribosome breakdown was
suggested as the source of GFAJ’s phosphate as evidenced by experimental work
with E. coli (Basturea et al. 2012).
Our refutation of these papers is that the efforts by Reaves et al. and Erb et al.
employed additional treatments to the extracted DNA before analyses. After an
initial phenol chloroform extraction, Reaves et al. purified the DNA with RNase A
and Proteinase K followed by phenol chloroform extraction and ethanol precipitation. Erb et al. did not use RNase A but repeated 70% ethanol rinses. The DNA was
then resuspended in water before it was analyzed. Hence any susceptible arsenateesters would have been broken down by the extraction process itself. We did not
employ such methods, relying on NanoSIMS and X-ray spectroscopy for our results.
Indeed, in our paper we made note of the instability of the molecule in water and
suggested that it could have been stabilized internally by such molecules as
poly-β-hydroxybutyrate. Moreover, Wolfe-Simon pointed out that DNA within
680
R. S. Oremland