SEPARATION AND CHARACTERIZATION OF STRUCTURAL ISOMERS OF
PERFLUORINATED COMPOUNDS
Hoon Yoo1, Paul D. Jones1, Patrick W. Bradley1, Miodrag Gužvić1,2, Brad L. Upham1,2, James E. Trosko1,2, John L Newsted3 and John P. Giesy1
Food Safety and Toxicology Center, Zoology Department and Center for Integrative Toxicology, 2Department of Pediatrics and Human
Development, Michigan State University, East Lansing, MI 48824, USA. 3ENTRIX Inc., Okemos, MI 48864, USA.
1National
RESULTS and DISCUSSION
ABSTRACT
have separated a commercial PFOS mixture into three separate isomer peaks (Fig.1) .
PFOS: Analytical chemistry
Inhibition of GJIC by various PFCs
14.43
100
The largest peak (peak III), presumed to be the linear isomer constitutes greater than
14.64
100
A
100
B
13.56
C
14.92
and 34% of the mass respectively. The mixture also contained traces of PFOA and
shorter chain sulfonates that were not quantified. Using a preparative HPLC column
%
13.74
(150mm x 21.2mm) the linear chain peak was separated from the branched chain to
greater than 99.5% purity (Fig. 2).
12.61
To assess the potential epigenetic toxicity of the purified isomers and other PFCs, a
gap junction intercellular communication (GJIC) assay was conducted. In the current
study, the linear PFOS isomer was approximately equipotent, on a mass basis, at
%
%
0
0.00
10.00
SIR of 7 Channels ES499.00 1.00Da
8.83e5
Area
12.50
13.95
Time
5.00
SIR of 7 Channels ES499.00 1.00Da
2.22e6
Area
SIR of 7 Channels ES499.00 1.00Da
4.42e6
Area
15.00 20.00 25.00
0
0.00
0
0.00
Time
5.00
10.00 15.00
20.00 25.00
Time
5.00
10.00 15.00 20.00 25.00
Figure 1. Chromatogram of technical PFOS mixture (A), purified linear PFOS (B),
and enriched branched chain PFOS (C).
CF3
CF3
INTRODUCTION
While current analytical methods are able to separate and quantify different PFCs such as
sulfonates from carboxylates and different carbon chain lengths, little attention has been
paid to the separation of branched from linear chain isomers and, in general standards
used for PFC quantification have not been characterized for their isomer distribution.
•
Similarly, all toxicological studies to date have been conducted using relatively crude
commercial available PFC preparations.
•
In this study, separation of isomers in the commercial mixture PFOS was achieved using
a HPLC-MS and a fraction collector. Potential epigenetic toxicity of purified PFOS and
various PFCs was evaluated using GJIC inhibition assay.
0.25
1
Probably 3 component mix
10
100
PFOS concentration (mg/L)
Table 1. NOEC and LOEC values of
PFCs in the SL/DT assay (mg/L).
Compounds
NOEC
Linear PFOS
< 3.125
LOEC
3.125
Branched PFOS
3.125
6.25
PFHA
3.125
6.25
PFHS
<3.125
3.125
PFOA
<3.125
3.125
TH-PFDA
6.25
12.5
1H,1H-PFUnA
TH-PFDoA
>100
>100
Figure 5. Gap junction inhibition at multiple concentrations of various PFCs.
All CF2
inhibit GJIC (Fig. 5 and 6).
•
0.50
B. Branched PFOS: 19F NMR
A. Linear PFOS: 19F NMR
similar potency for inhibition of GJIC. Among the PFCs investigated, PFHS, PFHA, PFOA,
and tetrahydro (TH) PFDA inhibited the GJIC, while TH-PFUnA, 1H,1H,PFDoA did not
0.75
0.00
inhibiting GJIC compared to the commercial mixture. The enriched branched and linear
isomers had similar potencies (Fig. 3). This indicates the different PFOS isomers have
PFHA
PFHS
PFOA
TH-PFDA
1.00
13.84
63 % of the materials mass while the earlier eluting peaks, peaks I and II constitute 2.5%
Fraction of Control
Using modifications of current perfluorinated compound (PFC) analytical methods we
PFHA
PFHS
PFOA
TH-PFDA
1H,1H-PFUnA
TH-PFDoA
Octet
Adjacent to
-SO3
F F F
F C C C F
F
F
D. Branched PFOS: Proton NMR
C. Linear PFOS: Proton NMR
CD3OD
CD3OD
CD3OD
CD3OD
CD3OD
CD3OD
Proton adjacent
to carbonyl ?
METHODS and MATERIALS
Preparative Chromatography
Preparative chromatography was conducted using a gradient of MeOH in 5mM ammonium
acetate. Separations were carried out on an HP 1100 HPLC system using a Betasil 5
150mmX21.2mm, 5um, C18 column (Thermo Electron, San Jose, CA). The effluent from
the column was split approximately 99:1 between a fraction collector and a Micromass
Platform II spectrometer running in SIR mode to monitor chromatographic separation.
Figure 2. 19F and proton spectra determined by NMR analysis
1.00
NMR Methods
Purified standards were dissolved in deuterated methanol and analyzed using both
proton (64 scans, 1 sec cycle delay, 499.96 MHz) and fluorine (32 scans, 5 sec cycle delay,
470.38 MHz) Varian Unity 500 MHz NMR.
Fraction of Control
Analytical chromatography was conducted using a 250mmX2mm, 5um, 100Å C18 column
(Thermo Electron). Solvent A was 10% MeOH in 5mM ammonium acetate, solvent B was
100% MeOH. Analytes were quantified by use of a HP 100 HPLC system interfaced to a
Micromass Platform II mass spectrometer. The mass spectrometer was run in
electrospray negative mode and analytes were determined by single ion monitoring (SIR).
CONCLUSIONS
PFOS: Toxicology – GJIC inhibition assay
*
Analytical Chromatography
Figure 6. Comparison of SL/DT assay at the greatest exposure (100 mg/L) of
various PFCs (refer to Fig.4 for the vehicle control).
*
Commercial
Linear PFOS
Branch PFOS
*
0.75
0.25
* = LOEC
0.00
1
10
Technical PFOS mixtures can be resolved in to at least 4 peaks.
•
Branched chain PFOS is ‘pure’ as determined by NMR (and >99% by LC/MS).
•
Branched chain contains at least 3 components – one appears to be a
terminal branch.
Figure 3. Gap junction
inhibition at multiple
concentrations of PFOS
isomers. Stars denote the
least observed effect
concentration (LOEC).
0.50
•
•
Branched chain preparations may contain some C-H bond impurities ?
•
Linear PFOS appears to be only slightly more potent than branched chain in
the GJIC assay.
•
Among the PFCs tested, PFOS was the most potent inhibitor of GJIC.
•
The effect of chain length on the GJIC assay was consistent with previous
100
studies. Only the compounds with C6-C10 inhibited GJIC in a concentration-
PFOS concentration (mg/L)
dependent manner, irrespective of whether they were carboxylates or
sulfonates.
The Scrape Loading and Dye Transfer (SL/DT) assay for GJIC in vitro measurement
I) Rat liver epithelial cells (2mL) +
20 uL of PFCs (15 min exposure)
The distance the Lucifer Yellow dye travels (within 5 min)
away from the scrape line is indicative of the level of GJIC
within a culture (Fig. 4 and 6)
Vehicle control
Linear PFOS
Branched PFOS
•
Longer chain PFCs (TH-PFDoA and 1H,1H-PFUnA) were inactive in the GJIC
assay at the greatest exposure.
II) Add the Lucifer Yellow (457 Da)
III) 3X scrapes with a surgical blade
IV) Observe under
fluorescent microscope
V) Analyze image
with Gel-Expert
Figure 4. Comparison of SL/DT assay at the greatest exposure (100 mg/L) of linear
and branched PFOS
FLUOROS Meeting
August 18-20, 2005,
Toronto, ON, CANADA