Characterization and origin of the ‘B’ and ‘C’ compounds in the acid/neutral forensic signatures of heroin – products from the acylation of porphyroxine and subsequent hydrolysis
John F. Casale,* Ellen S. Casale, Steven G. Toske, Patrick A. Hays and Sini Panicker

Two significant compounds often found in the gas chromatographic analysis of the acid/neutral extracts from illicit heroin have remained uncharacterized for 30 years. The unknown compounds are referred to as the ‘B’ and ‘C’ compounds. It has been postu- lated that these compounds arise from acetylation of porphyroxine, a rhoeadine alkaloid found at trace levels in the opium poppy, Papaver somniferum. Porphyroxine was isolated from opium and acetylated to produce N,O8-diacetylporphyroxine. Mild hydroly- sis produced N,O8-diacetyl-O14-desmethyl-epi-porphyroxine (the C compound) and N-acetyl-O14-desmethyl-epi-porphyroxine (the B compound). Both N,O8-diacetyl-O14-desmethyl-epi-porphyroxine and N-acetyl-O14-desmethyl-epi-porphyroxine were deter- mined to be C-14 epimers of porphyroxine and N,O8-diacetylporphyroxine. The non-epimerized isomers of the B and C com- pounds were also detected in illicit heroin, but at much lower levels. Chromatographic and spectroscopic data are presented for the aforementioned compounds. The presence/absence and relative concentrations of these compounds is presented for the four types of heroin (Southwest Asian, South American, Southeast Asian, and Mexican). The prevalence of detection for the B and C compounds are Southwest Asian = 92-93%, South American = 64-72%, Southeast Asian = 45-49%, and Mexican ≤ 3%. When
detected, the overall trend of relative concentrations of dicaetylporhyroxine, the B-compound, and C-compound is Southwest Asian > South American > Southeast Asian, each by an order of magnitude. These compounds were rarely detected in Mexican heroin. The presence/absence and relative concentrations of these compounds provide pertinent forensic signature characteris- tics that significantly enhance the final regional classifications. Copyright © 2015 John Wiley & Sons, Ltd.
Additional supporting information may be found in the online version of this article at the publisher’s web site.

Keywords: porphyroxine; gas chromatography-mass spectrometry; heroin signature; heroin profiling; forensic science

Introduction
Gas chromatographic analysis of the acid/neutral extracts from illicit heroin has been one of three essential chromatographic tech- niques for identifying the type/source of heroin, and has been uti- lized for over 30 years.[1–10] This methodology examines byproducts resulting from the acetylation of crude morphine, and has recently been enhanced to include mass spectral detection and the addition of newly identified target compounds.[1,2] However, two significant and major target compounds have remained uncharacterized since the advent of this analysis; their opium alkaloid origin was also un- known. The unknown compounds are referred to as the ‘B’ and ‘C’ compounds. The presence/absence and relative amounts of these compounds gives evidence to how heroin was produced from mor- phine as well as some insight to the origin of the opium. This laboratory’s previous (unpublished) work on the isolation of these compounds from crude heroin was unsuccessful, but indicated that both compounds had structural features similar to porphyroxine (Figure 1).
Porphyroxine (1), sometimes referred to as Papaverrubine-D, is a rhoeadine alkaloid found at trace levels in the opium poppy,

Papaver somniferum. It was first described and crudely isolated over 170 years ago by Merck,[11] and first isolated in a pure state by Genest and Farmilo;[12] its structure was elucidated in the 1960s via nuclear magnetic resonance spectroscopy (NMR).[13] Since the mid-1800s, rhoeadines and papaverrubines have been associated with opium characterization using strong acids, and porphyroxine has been employed forensically since the 1950s as an identification marker for raw opium. When treated with acids, raw opium contain- ing porphyroxine generates a red colour due to a red iminium salt formed from acid catalyzed reaction and molecular rearrangement.[14] Early publications have reported that rhoeadines and papaverrubines are prevalent in many Papaver varieties such as
P. glaucum, P. gracile, and P. decaisnei, etc., but only traces are

* Correspondence to: John Casale, Special Testing and Research Laboratory, US Drug Enforcement Administration, 22624 Dulles Summit Court, Dulles, VA 20166-9509, USA.
E-mail: [email protected]

Special Testing and Research Laboratory, U.S. Drug Enforcement Administration, 22624 Dulles Summit Court, Dulles, VA, 20166-9509, USA

Figure 1. Structural formulae and reaction pathways.

identified in P. somniferum and P. setigerum. This laboratory has conducted hundreds of porphyroxine colour tests on raw opium samples. Despite its low content, the porphyroxine test with hydro- chloric acid is positive for most raw opium samples, especially for those collected in Afghanistan. The slightly red colour of opium latex oozing out of Afghanistan poppy pods immediately after lanc- ing is believed to be caused by porphyroxine.
Porphyroxine is a tetracyclic alkaloid having a phenolic OH and methoxy-substitution on the A-ring, a secondary amine within the B-ring, a methylenedioxy-substitution on the C-ring, and a cyclic acetal moiety within the D-ring. The phenolic OH causes it to be co-extracted with morphine during the crude illicit processing of opium. It is readily acetylated to form a neutral amide during the production of heroin from morphine.
This work describes the isolation of porphyroxine 1 from opium, its acetylation to diacetylporphyroxine 2, and the subsequent acidic hydrolysis to the C compound 3 and the B compound 4 (Figure 1). Chromatographic and spectroscopic data are presented which characterize 2, 3, and 4. C-14 epimerization had occurred in the formation of 3 and 4. The non-epimerized isomers 5 and 6 of 3 and 4, respectively, which are also detected in illicit heroin, were also characterized. Compounds 2–6 are all commonly detected in illicit Southwest Asian heroin, but less prominently or not at all in other types of heroin. The presence/absence and relative con- centrations of these compounds is discussed for the four main types of heroin (Southwest Asian, Southeast Asian, South American, and Mexican), based on the analyses of over 900 individual samples.

Experimental
Opium and heroin samples
Opium seized in Afghanistan (1.876 kg) was utilized for the isolation of porphyroxine and was stored in a vault under ambient condi- tions prior to workup. The concentration of porphyroxine in the opium was estimated to be less than 0.1 w/w %. Authentic samples of South American (SA), Mexican (MEX), Southwest Asian (SWA), and Southeast Asian (SEA) heroin were selected from this laboratory’s authentic reference materials collection.

Materials
Aluminum oxide (activated, basic, Brockmann I, 150 mesh), silica gel (70–230 mesh), and all solvents/reagents (reagent grade or better) were obtained from Sigma Aldrich (St Louis, MO, USA). N-Methyl- N-trimethylsilyltrifluoro-acetamide (MSTFA) was purchased from Thermo Scientific (Waltham, MA, USA) in 1 mL sealed glass ampules.

Gas chromatography-mass spectrometry (GC-MS)
Two GC-MS systems were employed for this study. System #1: GC- MS analyses were performed using an Agilent Model 5975C quad- rupole mass-selective detector (MSD) interfaced with an Agilent 7890 gas chromatograph (Wilmington, DE, USA). The MSD was op- erated in the electron ionization mode (EI) with an ionization poten- tial of 70 eV, a scan range of 34–700 mass units, at 1.34 scans/s. The GC system was fitted with a 30 m×0.25 mm ID fused-silica capillary column coated with DB-1 (0.25 μm) (J & W Scientific, Rancho Cordova, CA, USA). The oven temperature was programmed as fol- lows: Initial temperature, 100 °C; initial hold, 0.0 min; program rate, 6 °C/min; final temperature, 300 °C; final hold, 15.67 min (total run time, 39.0 min). The injector was operated in the split mode (21:1) and at a temperature of 280 °C. The auxiliary transfer line to the MSD was operated at 280 °C. Injection volumes were 2 μL. System #2: GC-MS analysis was performed using a Thermo Electron Trace gas chromatograph outfitted with a PTV injection port and coupled with a Thermo Electron Polaris-Q ion trap mass spectrometer. All GC-MS parameters were identical to those previously reported.[1]

Nuclear magnetic resonance (NMR)
NMR spectra were obtained using an Agilent VNMRS 600 MHz NMR with a 5 mm broadband pulse field gradient probe (Palo Alto, CA, USA). The sample temperature was maintained at 25 °C. Standard Agilent pulse sequences were used to collect the following spectra: Proton, carbon-13 (proton decoupled), 1D-NOESY (spatial nearness
spectra showing which protons are spatially near (<4 Å) the selec- tively excited proton, 500 ms mixing time used), and gradient ver-
sions of the 2-dimensional experiments COSY (proton to proton coupled correlations), HSQC (directly bonded protons to carbon correlations), and HMBC (long range 2, 3, and 4 bond proton to car- bon correlations). Samples were dissolved in deuterated chloro- form (CDCl3) containing 0.03% v/v tetramethylsilane (TMS, 0 ppm reference) (Cambridge Isotopes, Tewksbury, MA, USA). Data pro- cessing and structure elucidation were performed using Agilent NMR software and Applied Chemistry Development Structure Eluci- dator software (ACD/Labs, Toronto, Canada).

Isolation of porphyroxine 1
The opium was extracted in four separate portions. The typical pro- cedure is as follows: Approximately 450 grams of opium was stirred overnight in 5 L hot water (60 °C) and allowed to cool to room tem- perature. The solution was centrifuged (1200 g) to remove insoluble materials; the supernatant was extracted with diethyl ether (2×1800 mL), followed by methylene chloride (2×2000 mL). These extracts were discarded. The aqueous phase was extracted with CHCl3 (2×2000 mL) and saved. The aqueous phase was adjusted to pH 9-10 with 3 M NaOH and extracted again with chloroform (2×2000 mL). All chloroform extracts were combined, dried over an- hydrous sodium sulfate, filtered, and evaporated in vacuo to a dark viscous oil (ca. 40 grams/batch).

Characterization and origin of the “B” and “C” compounds in forensic signatures of heroin

and Analysis

Alumina chromatography for the preparative isolation of porphyroxine was performed in four separate portions. Approxi- mately 40 g of extract was loaded onto 800 g of basic alumina (adjusted to 4% water w/w) and eluted with chloroform (500 mL), chloroform/acetone (85:15, 500 mL), chloroform/acetone (1:1, 500 mL), acetone (500 mL), acetone/methanol (9:1, 500 mL), acetone/methanol (1:1, 500 mL), and methanol (1 L). Fractions were monitored via GC-MS (system #1), and those containing enriched porphyroxine (chloroform/acetone) were combined and evaporated in vacuo to a dark resinous oil (ca. 2 g/column).
Porphyroxine was further purified in four portions via silica gel chromatography. Approximately 2 g of enriched porphyroxine was chromatographed on 30 g of silica gel using 100 mL portions of the following mixtures of isopropanol/ammoniated-hexane: 8:2, 7:3, 6:4, and 1:1. Fractions were monitored via GC-MS (system #1) and those containing porphyroxine (≥98%) were combined and evaporated in vacuo to an off-white powder. Fractions containing 85-97% porphyroxine were evaporated in vacuo and recrystallized from isopropanol/hexane (1:3) until the chromatographic purity was ≥98%. The total recovery of porphyroxine was 159 mg (0.008% w/w) of white crystalline material. Exposure to air causes the material to slowly turn to a light tan colour. 1H NMR (600 MHz, CDCl3, coupling constants, J, are in Hz) δ 7.41 (s, 1H), 7.06 (dd, J = 8.3,
J < 1, 1H), 6.81 (d, J = 8.3, 1H), 6.65 (s,1H), 6.07 (d, J = 1.4,1H), 5.92
(d, J = 1.4, 1H), 5.83 (s, 1H), 4.68 (d, J = 8.8, 1H), 3.89 (s, 3H), 3.74
(d, J = 8.8, 1H), 3.68 (s, 3H), 3.32 (ddd, J = 13.3, 5.1, 2.0, 1H), 2.89
(ddd, J = 14.6, 11.0, 2.0, 1H), 2.81 (bdd, J = 13.3, 11.0, 1H), 2.73 (bdd, J = 14.6, 5.1, 1H); 13C NMR (150 MHz, CDCl3) δ 146.6, 144.8, 143.8, 143.4, 134.2, 132.9, 130.5, 121.0, 118.0, 112.0, 111.6, 109.0,
101.5, 98.2, 79.0, 58.3, 56.1, 55.3, 48.0, 40.5; GC-MS (EI) m/z (rel.
intensity) 179 (M+-192, 100), 356 (M+-15, 75), 371 (M+., 53).

Acetylation of porphyroxine to N,O8-diacetylporphyroxine 2
Porphyroxine (17.0 mg, 0.046 mmol) was heated with 800 μL of acetic anhydride (8.46 mmol) in a sealed glass tube for one hour at 105 °C. Excess acetic anhydride was evaporated in vacuo to ob-
tain 20.0 mg (96%) of light tan powder. The chromatographic purity was >98% and the material was suitable for GC-MS and NMR anal- yses without further purification. Major rotamer, 1H NMR (600 MHz, CDCl3) δ 7.35 (s, 1H), 6.82 (d, J = 8, 1H), 6.73 (s, 1H), 6.63 (bd, J = 8, 1H),
6.12 (d, J = 1.2, 1H), 6.06 (s, 1H), 6.01 (d, J = 1.2, 1H), 4.88 (d, J = 9.1,
1H), 4.70 (d, J = 9.1, 1H), 4.39 (m, 1H), 3.81 (s, 3H), 3.70 (s, 3H), 3.28
(m, 1H), 2.98 (bdd, J = 16.0, 9.0, 1H), 2.80 (dt, J = 14.1, 9.0, 1H), 2.32
(s, 3H), 1.90 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 171.7, 169.1, 150.3,
148.0, 145.5, 138.1, 133.0, 129.2, 128.4, 118.1, 116.9, 116.3, 114.6,
108.8, 102.1, 98.6, 75.1, 61.8, 55.9, 55.6, 37.0, 32.4, 21.1, 20.7.
Minor rotamer, 1H NMR (600 MHz, CDCl3) δ 7.37 (s, 1H), 6.77 (d, J = 8.1, 1H), 6.71 (s, 1H), 6.52 (bd, J = 8.1, 1H), 6.13 (s, 1H), 6.09
(d, J = 1.2, 1H), 5.98 (d, J = 1.2, 1H), 5.47 (d, J = 10, 1H), 4.82 (d, J = 10,
1H), 3.81 (s, 3H), 3.78 (m , 1H), 3.57 (s, 3H), 3.31 (m, 1H), 3.15 (m,
1H), 3.15 (m, 1H), 2.22 (s, 3H), 2.15 (s, 3H); 13C NMR (150 MHz, CDCl3)
δ 171.4, 166.4, 150.3, 147.4, 145.0, 138.7, 131.3, 130.0, 129.4, 119.0,
117.5, 116.5, 114.1, 108.7, 101.9, 98.3, 74.8, 57.4, 56.1, 54.0, 40.3,
33.2, 22.2, 21.2; GC-MS (EI) m/z (rel. intensity) 412 (M+-43, 100),
179 (M+-276, 63), 455 (M+., 7).

Hydrolysis of N,O8-diacetylporphyroxine 2 to N,O8-diacetyl-O14- desmethyl-epi-porphyroxine (C-compound) 3
Porphyroxine (24.5 mg, 0.066 mmol) was heated with 800 μL of acetic anhydride (8.46 mmol) in a sealed glass tube for one hour

at 105 °C. After cooling, water (5.0 mL) was added to quench unreacted acetic anhydride. The solution was allowed to stand at room temperature for 8 days. The reaction was then extracted with CHCl3 (2×5 mL), and the extracts were combined, washed with 1 N H2SO4 (3 mL), dried over anhydrous Na2SO4, and evaporated to dryness to obtain 23.4 mg (80%) of tan powder.
Major rotamer, 1H NMR (600 MHz, CDCl3) δ 7.25 (s, 1H), 6.82 (d, J = 8.0, 1H), 6.74 (s, 1H), 6.65 (dd, J = 8.0, 1.0, 1H), 6.35 (s, 1H),
6.09 (d, J = 1.4, 1H), 5.98 (d, J = 1.4, 1H), 5.36 (d, J = 9.6, 1H), 4.48
(d, J = 9.6, 1H), 4.35 (m, 1H), 3.80 (s, 3H), 3.28 (m, 1H), 3.03 (m, 1H),
2.88 (dt, J = 14.3, 8.6, 1H), 2.32 (s, 3H), 1.88 (s, 3H); 13C NMR
(150 MHz, CDCl3) δ 172.0, 169.2, 150.2, 147.9, 145.4, 138.1, 133.2,
128.5, 128.2, 118.0, 117.4, 117.2, 114.7, 108.7, 101.9, 89.8, 69.1,
62.3, 56.1, 37.2, 32.4, 21.0, 20.7.
Minor rotamer, 1H NMR (600 MHz, CDCl3) δ 7.27 (s, 1H), 6.76 (m, 2H), 6.52 (bd, J = 8.0, 1H), 6.32 (s, 1H), 6.06 (bs, 1H), 6.00 (d, J = 1.3,
1H), 5.31 (d, J = 10.0, 1H), 5.29 (d, J = 10.0, 1H), 3.80 (s, 3H), 3.77
(m, 1H), 3.40 (dt, J = 15.9, 9.0, 1H), 3.18 (m, 1H), 3.18 (m, 1H), 2.30 (s, 3H), 2.14 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 172.1, 169.2,
~150, 147.3, 144.9, 138.7, 131.6, 129.6, 128.5, 118.9, 117.8, 117.5,
114.7, 108.7, 102.3, 89.8, 68.7, 57.8, 55.9, 40.5, 33.2, 21.2, ~21; GC-MS
(EI) m/z (rel. intensity) 470 (M+-43 + TMS, 100), 43 (M+-398, 64),
513 (M+. + TMS, 8).

Hydrolysis N,O8-diacetyl-O14-desmethyl-epi-porphyroxine 3 to N-acetyl-O14-desmethyl-epi-porphyroxine (B-compound) 4
Compound 3 (23.4 mg, 0.053 mmol) was dissolved into acetonitrile (1.0 mL), diluted with dilute acetic acid (6.0 mL, 3.5 M), and heated at 75 °C for 5 days. The reaction was diluted with water (7.0 mL), ex- tracted with CHCl3 (2×10 mL), and the extracts were combined, washed with 1 N H2SO4 (3 mL), dried over anhydrous Na2SO4, and evaporated to dryness to obtain 17.4 mg (82%) of reddish powder. Major rotamer, 1H NMR (600 MHz, CDCl3) δ 7.13 (s, 1H), 6.81 (d, J = 8.0, 1H), 6.65 (dd, J = 8.0, 1.2, 1H), 6.63 (s, 1H), 6.37 (s, 1H),
6.11 (d, J = 1.4, 1H), 6.02 (d, J = 1.4, 1H), 5.31 (d, J = 9.4, 1H), 4.46
(d, J = 9.4, 1H), 4.35 (m, 1H), 3.85 (s, 3H), 3.19 (m, 1H), 2.98 (bdd,
J = 15.4, 8.8, 1H), 2.83 (m, 1H), 1.87 (s, 3H); 13C NMR (150 MHz, CDCl3)
δ 172.0, 147.8, 145.7, 145.4, 144.1, 128.9, 128.3, 125.9, 117.5, 117.2,
113.1, 109.7, 108.7, 102.1, 89.8, 69.3, 62.2, 55.9, 37.3, 31.9, 21.0.
Minor rotamer, 1H NMR (600 MHz, CDCl3) δ 7.14 (s, 1H), 6.75 (d, J = 8.0, 1H), 6.60 (s, 1H), 6.52 (dd, J = 8.0, 1.0, 1H), 6.33 (s, 1H),
6.08 (d, J = 1.4, 1H), 5.99 (d, J = 1.4, 1H), 5.28, (obscured, 1H), 4.50
(d, J = 9.5, 1H), 3.86 (s, 3H), 3.75 (m, 1H), 3.34 (dt, J = 15.6, 9.1, 1H),
3.15 (m, 1H), 3.06 (m, 1H), 2.13 (s, 3H); 13C NMR (150 MHz, CDCl3) δ
172.1, 147.2, 145.8, 144.9, 144.7, 130.0, 128.6, 124.2, 117.8, 117.5,
112.6, 110.8, 108.6, 101.9, 89.8, 69.0, 61.8, 56.1, 40.6, 32.7, 21.1; GC-
MS (EI) m/z (rel. intensity) 73 (TMS, 100), 500 (M+-43 + 2TMS, 29),
543 (M+. + 2TMS, 8).

Results and discussion
Supplemental material may be found in the online version of this paper. The ring numbering assignments used by Brochmann- Hanssen and Hirai were utilized within the text of this work, and IUPAC numbering was utilized in the Supplemental NMR data tables.

Isolation of porphyroxine 1
This laboratory’s previous (unpublished) work on the isolation of the B and C compounds from crude heroin was unsuccessful due

to isomerization of both compounds in the acidic conditions uti- lized during the workups. Hence, an essentially neutral pH extrac- tion was employed to remove papaverine and noscapine from opium. Adjusting the pH to strongly basic with NaOH allowed for the extraction of the remaining alkaloids, except morphine, which remained in solution as sodium morphinate. Alumina column chromatography of these latter extracts gave enriched fractions (0.5–2%) of 1 that contained codeine as the principle contaminant. Silica gel column chromatography isolated 1 in the early fractions while codeine eluted significantly later. Purified 1 gave MS (Figure 2a) fully consistent with NIST data. The 1H NMR spectrum was identical to published literature after converting the published Tau, τ, chem- ical shift data to the current Delta, δ, chemical shift system.[13] The relative stereochemistry for the D-ring also matched the published

work, with a coupling constant (J = 8.8 Hz) for the methine protons located on C-1 and C-2. This result is consistent for a vicinal trans- diaxial proton arrangement. Also, the presence of a nuclear Overhauser effect (NOE) correlation from the proton on C-1 to the proton on C-14 using a 1D-NOESY experiment proved the methoxy group was equatorial on the anomeric C-14. In addition, the two- dimensional NMR data was consistent for the porphyroxine structure.

Synthesis and characterization of N,O8-diacetylporphyroxine 2
Acetylation of 1 produced 2 in high yield and purity. The mass spectrum for 2 (Figure 2b) produced a base peak at m/z 412 and a moderate intensity molecular ion at m/z 455; the spectrum and

Figure 2. Electron ionization mass spectrum of (a) porphyroxine 1, and (b) diacetylporphyroxine 2.

retention time were identical to a prominent uncharacterized and untargeted peak found in the acid/neutral extracts of SWA her- oin signatures (Figure 3a, peak #5). See Table 1 for peak identifica- tions. Due to its amide properties, 2 existed as two rotameric forms (2:1 ratio) in solution during NMR analyses; its proton and carbon spectra were determined to be consistent with diacetylporphyroxine. The relative stereochemistry for the D-ring was also a match to 1, based on a similar coupling constant (J = 9.1 Hz) for the methine protons located on C-1 and C-2. Also, the NOE was intact for the 1,3-diaxial protons located on C-1 and C-14, which confirmed that the relative stereochemistry for the D-ring was preserved.

Figure 3. Partial reconstructed total ion chromatograms of acid/neutral extracts of typical (a) Southwest Asian heroin, (b) South American heroin,
(c) Southeast Asian heroin, and (d) Mexican heroin. See Table 1 for peak identification.

Synthesis and characterization of N,O8-diacetyl-O14-desmethyl- epi-porphyroxine (C Compound) 3

Mild acidic hydrolysis of 2 provided 3 (C compound) in good yield and purity. It could not be chromatographed underivatized via GC- MS, but its spectrum was readily visualized as the mono-trimethylsilyl derivative (Figure 4a). The mass spectrum of derivatized 3 gave a base peak at m/z 470 and a moderate molecular ion at m/z 513; the spectrum and retention time were identical to the derivatized C compound, found in the acid/neutral extracts of heroin signatures (Figures 3a–3c, peak #2). The amide properties of 3 also gave two rotameric forms (2:1) during NMR analyses; the spectra were

consistent with hydrolysis/elimination of the methoxy-methyl substituent at C-14 on the cyclic hemiacetal moiety (D-ring), and epimerization at C-14 of diacetylporphyroxine. No NOE effect was observed between the C-1 and C-14 protons when each was selectively excited. This inferred that a selective epimerization had occurred during the slow hydrolysis to the hemiacetal 3, and the resulting hydroxyl group was located in the axial position on C-14. This result is consistent with cyclic acetal reactions, where an anomeric effect favors axial nucleophilic addition products.[15] As expected, the coupling constant (J = 9.6 Hz) for the C-1 and C-2 protons was consistent with a trans-diaxial arrangement.

Figure 4. Electron ionization mass spectrum of (a) N,O8-Diacetyl-O14-desmethyl-epi-porphyroxine 3 as the trimethylsilyl derivative, and (b) N-Acetyl-O14- desmethyl-epi-porphyroxine 4 as the bistrimethylsilyl derivative.

Synthesis and characterization of N-acetyl-O14-desmethyl-epi- porphyroxine (B-compound) 4
Rigorous hydrolysis conditions were required for the formation of 4. Similar to compound 3, 4 could not be chromatographed underivatized via GC-MS, but its spectrum was readily visualized as the bis-trimethylsilyl derivative (Figure 4b). The mass spectrum of derivatized 4 gave a base peak at m/z 73 and a moderate mo- lecular ion at m/z 543; the spectrum and retention time were identical to the derivatized B compound, found in the acid/ neutral extracts of heroin signatures (Figures 3a–3c, peak #1). The amide interconversion of 4 also gave two rotameric forms (2:1) during NMR analyses; the spectra were consistent with hy- drolysis or removal of the O8-acetyl and retaining the same ste- reochemistry as 3 at C-14.

N,O8-Diacetyl-O14-desmethylporphyroxine 5 and N-acetyl-O14-
desmethylporphyroxine 6
During the synthesis of 3 and 4, small amounts of the C-14 non- epimerized isomers 5 and 6 (Figure 5) were detected at relative ra- tios of approximately 13:1 (non-epimerized compounds = 1). Each gave virtually identical mass spectra to its respective epi-isomer (Figures 4a–4b), but eluted later in the chromatographic profile (Figure 3a, peaks 3 and 4). The relative ratios of 3 vs. 5 and 4 vs. 6 during synthesis were approximately the same (ca. 13:1) as seen in Figure 3a.

Forensic analysis of heroin samples
The identity of the targeted unknown, C, and B compounds (2, 3, and 4) are now known. In addition, two previously untargeted iso- mers of the B and C compounds have been characterized (5 and 6). All five compounds arise from the acetylation of porphyroxine contaminated-morphine, during illicit production of heroin. As de- termined experimentally, the relationship for the formation of these compounds is 1 → 2 → 3 + 5 → 4 + 6. The average concentration and percent of samples detected for 2, 3, and 4 found in the four types of heroin are given in Table 2 for over 900 heroin samples. Since 5 and 6 are newly identified targets, archived data does not exist for them. Typical partially reconstructed total ion chromato- grams for each major type of heroin are illustrated in Figure 3. As seen in Table 2 and Figure 3, SWA heroin contains the greatest con- centration of porphyroxine-related compounds (1-2 orders of mag- nitude greater), SA heroin contains the next highest levels of these compounds, followed by SEA heroin, and then rarely detected in MEX heroin (1–3% of MEX samples). Quenching a hot acetylation reaction with water will accelerate the conversion of 2 → 3 + 5 and 4 + 6. Additionally, transformation of heroin base into heroin HCl in hot organic solvents (the typical SA process) will accelerate

Figure 5. Structures of N,O8-Diacetyl-O14-desmethylporphyroxine 5 and N-Acetyl-O14-desmethylporphyroxine 6.

these conversions. The prevalence of detection for compounds 3
and 4 are SWA = 92–93%, SA = 64–72%, SEA = 45–49%, and
MEX ≤ 3%. When detected, the overall trend of relative concentra- tions of 2, 3, and 4 is SWA > SA > SEA > MEX, each by an order of magnitude. The presence/absence and relative concentrations of
these compounds provide pertinent forensic signature characteris- tics that significantly enhance the final regional classifications. The presence and relative concentrations of porphyroxine-related compounds in heroin may be directly related to origin; poppies grown in these regions would be expected to have distinct differ- ences in porphyroxine content. Currently, we are investigating the porphyroxine content of opium from the four heroin producing regions.
Despite the lack of current information on the abundance of 1 in the opium of various regions, the data indicates that Southwest Asian opium (primarily, from Afghanistan) contains a greater con- centration of 1. The most crudely manufactured heroin samples from Afghanistan display the highest amounts of 3, suggesting that the mild hydrolysis of 2 to 3 is routine in morphine to heroin processing in that region. Subsequent rigorous hydrolysis also takes place during the processing, but to a lesser extent. As the refinement process continues for heroin, the amounts of all porphyroxine-related impurities diminish in the final product, as ex- pected. Interestingly, despite its very crude manufacturing method, Mexican black tar and brown powder heroin do not show any mea- surable amounts of these impurities. Approximately 99% of ana- lyzed Mexican heroin does not contain diacetylporphyroxine 2. Since black tar is generated with open-pot cooking of morphine with acetic anhydride with no additional purification or precipita- tion steps, if porphyroxine is abundant in Mexican opium, diacetylporphyroxine would be expected to be present. The ab- sence of 2 in Mexican heroin could be due to (1) extremely low levels of 1 in Mexican opium; or more likely, or (2) due to the morphine-to-morphine HCl conversion step (also performed in SEA heroin processing) that is not done in SA or SWA heroin pro- cessing. Conversion to morphine HCl with concentrated HCl will likely convert porphyroxine to the previously noted iminium salt through an acid catalyzed reaction and molecular rearrangement, thus eliminating the possibility of the formation of 2–6.

Acknowledgement
The authors are indebted to Sam D. Cooper of this laboratory for his assistance in the production of Figure 3.

References
[1] D. R. Morello, S. D. Cooper, S. Panicker, J. F. Casale. Signature profiling and classification of illicit heroin by GC-MS analysis of acidic and neutral manufacturing impurities. J. Forensic Sci. 2010, 55, 42.
[2] S. G. Toske, S. D. Cooper, D. R. Morello, P. A. Hays, J. F. Casale, E. Casale. Neutral heroin impurities from tetrahydrobenzylisoquinoline alkaloids.
J. Forensic Sci. 2006, 51, 308.
[3] B. A. Perillo, R. F. X. Klein, E. S. Franzosa. Recent advances by the U.S. Drug Enforcement Administration in drug signature and comparative analysis. Forensic Sci. Int. 1994, 69, 1.
[4] R. B. Myors, P. T. Crisp, S. V. Skopec, R. J. Wells. Investigation of heroin profiling using trace organic impurities. Analyst. 2001, 126, 679.
[5] F. Besacier, C. H. Thozet, R. M. Tsangaris, J. Girard, A. Lamotte. Comparative chemical analyses of drug samples: general approach and application to heroin. Forensic Sci. Int. 1997, 85, 113.
[6] M. Collins, E. Casale, D. Hibbert, S. Panicker, J. Robertson, S. Vujic. Chemical profiling of heroin recovered from the North Korean merchant vessel Pong Su. J. Forensic Sci. 2006, 51, 597.
[7] L. Stromberg, L. Lunderberg, H. Neumann, B. Bobon, H. Huizer,
N. W. van der Stelt. Heroin impurity profiling: a harmonization study for retrospective comparisons. Forensic Sci. Inter. 2000, 114, 67.
[8] R. Dams, T. Benijts, W. E. Lambert, D. L. Massart, A. P. De Lenheer. Heroin impurity profiling: trends throughout a decade of experimenting. Forensic Sci. Int. 2001, 123, 81.

[9] A. Johnston, L. A. King. Heroin profiling: predicting the country of origin of seized heroin. Forensic Sci. Int. 1988, 95, 47.
[10] C. Barnfield, S. Burns, D. L. Byrom, A. V. Kemmenoe. The routine profiling of forensic heroin samples. Forensic Sci. Int. 1988, 39, 107.
[11] E. Merck. Ueber bengalisches Opium und einen bei Untersuchung desselben aufgefundenen eigenthüemlichen Stoff. Ann. Pharm. 1837, 21, 201.
[12] K. Genest, C. G. Farmilo. The preparation of porphyroxine from opium.
J. Pharm. Pharmacol. 1963, 15, 197.
[13] E. Brochmann-Hanssen, K. Hirai. Opium alkaloids V. Structure of porphyroxine. J. Pharm. Sci. 1967, 56, 1658.
[14] C. T. Montgomery, B. K. Cassels, M. Shamma. The rhoeadine alkaloids.
J. Nat. Prod. 1983, 46, 441.
[15] R. U. Lemieux. Effects of unshared pairs of electrons and their solvation on conformational equilibria. Pure Appl. Chem. 1971, 25, 527.

Supporting information
Additional supporting information may be found in the online ver- sion of this article at the publisher’s web site.Compound C