Psychedelics and the Human Receptorome

We currently understand the mental effects of psychedelics to be caused by agonism or partial agonism of 5-HT2A (and possibly 5-HT2C) receptors, and we understand that psychedelic drugs, especially phenylalkylamines, are fairly selective for these two receptors. This manuscript is a reference work on the receptor affinity pharmacology of psychedelic drugs. New data is presented on the affinity of twenty-five psychedelic drugs at fifty-one receptors, transporters, and ion channels, assayed by the National Institute of Mental Health – Psychoactive Drug Screening Program (NIMH-PDSP). In addition, comparable data gathered from the literature on ten additional drugs is also presented (mostly assayed by the NIMH-PDSP). A new method is introduced for normalizing affinity (Ki) data that factors out potency so that the multi-receptor affinity profiles of different drugs can be directly compared and contrasted. The method is then used to compare the thirty-five drugs in graphical and tabular form. It is shown that psychedelic drugs, especially phenylalkylamines, are not as selective as generally believed, interacting with forty-two of forty-nine broadly assayed sites. The thirty-five drugs of the study have very diverse patterns of interaction with different classes of receptors, emphasizing eighteen different receptors. This diversity of receptor interaction may underlie the qualitative diversity of these drugs. It should be possible to use this diverse set of drugs as probes into the roles played by the various receptor systems in the human mind.


Introduction
We currently understand the mental effects of psychedelics to be caused by agonism or partial agonism of 5-HT 2A (and possibly 5-HT 2C ) receptors (serotonin-2A and serotonin-2C receptors) [1]. This understanding was first developed in the 1980s [2][3][4] and has since been confirmed by a large body of evidence, as reviewed recently by Nichols [1]. However, many authors have commented that other receptors may also play a role [1,3,[5][6][7][8][9]. In this post-genome era of high-throughput assays, it is time to take a broader view, move beyond the common-denominator approach [6], and begin to explore the role of other receptors in shaping the mental effects of psychedelics, especially the qualitative differences among them.
The objective of this paper is to present the receptor binding profiles of the thirty-five drugs (Fig. 1, Fig. 2) of this study in such a way that they can be easily compared in both their similarities and their differences. This is intended to serve as a reference work on the multi-receptor affinity pharmacology of psychedelic drugs. The tables and figures are the heart of this manuscript. Some of them have been included as ''supporting information,'' because they exceed the size limits of standard tables and figures. However, this supporting information is no less central to the manuscript than the standard tables and figures.
Receptor data for ibogaine (Table S1), morphine (Table 1) and THC ( Table 2) was collected from a variety of sources. While ibogaine has been assayed at a wide variety of receptors, morphine and THC have not, so their data should be used with caution. Although morphine is not considered to be a psychedelic, and ibogaine, THC, and salvinorin A are not considered to be ''classical hallucinogens,'' these four compounds are included because they provide insights into additional receptor systems (salvinorin A -k (kappa opioid receptor), ibogaine -s (sigma receptor) and k, THC -CB (cannabinoid receptor), morphine -m (mu opioid receptor)). These additional compounds could also be thought of as active controls, as compared to the three presumably inactive controls of Fig. 1.

New PDSP Binding Assays
For this study, the NIMH-PDSP (http://pdsp.med.unc.edu/) has assayed sixteen phenylalkylamines, eight tryptamines and one ergoline (twenty-two psychedelics and three controls, Fig. 1) against a panel of fifty-one receptors, transporters, and ion channels. The methodology has been described previously by Glennon et al. [12]. Each compound is initially assayed at 10 mM against each receptor, transporter or ion channel (primary assay). Those that induce .50% inhibition (''hit'') are then assayed at 1, 10, 100, 1,000, and 10,000 nM to determine K i values (secondary assay). Each K i value (equilibrium dissociation constant, concentration at which 50% of the hot ligand is displaced by the test ligand) is calculated from at least three replicated assays. Details of how individual assays were conducted can be found at the NIMH-PDSP web site: http://pdsp.med.unc.edu/pdspw/binding.php. Table S2 shows raw K i data for the current study combined with data collected from the literature for the ten additional compounds; a total of thirty-five drugs and sixty-seven receptors, transporters and ion channels which were assayed. The table has been divided into three sections.
The second section displays seven sites at which most compounds were assayed, but at which there were no hits: 5ht3 (serotonin-3 receptor), H3 (histamine-3 receptor), H4 (histamine-4 Figure 1. Twenty-five drugs assayed for this study by the NIMH-PDPS. Twenty-five drugs assayed for this study by the NIMH-PDPS against fifty-one receptors, transporters and ion-channels. The twenty-five drugs include sixteen phenylalkylamines, eight tryptamines, and one ergoline. The three control drugs on the right include one representative from each structural class, and are believed to be non-psychedelic. doi:10.1371/journal.pone.0009019.g001 receptor), V1 (vasopressin-1 receptor), V2 (vasopressin-2 receptor), V3 (vasopressin-3 receptor), GabaA (GABA-A receptor).

Activity Assays
For the twenty-five compounds of Fig. 1, the NIMH-PDSP also performed activity assays at 5-HT 2A and 5-HT 2C . The E max values (maximal activity) are relative to 5-HT (serotonin), measuring Ca++ mobilization. Ca++ flux assays were performed using a FLIPRTETRA. The activity assays were conducted with cell lines which have very high receptor expression levels (e.g. plenty of 'spare receptors'). Under such conditions partial agonists will have considerable agonist activity. The data represent the mean 6 variance of computer-derived estimates from single experiments done in quadruplicate. Thus, the four observations are averaged and a single estimate with error is provided (Table S3).

Sources
The following compounds ( Fig. 1) were used in the study:  . Ten drugs whose receptor profiles were collected from the literature. Ten drugs whose receptor profiles were collected from the literature. All but ibogaine, THC, and morphine were assayed by the NIMH-PDSP. doi:10.1371/journal.pone.0009019.g002

Normalization
The raw K i values are distributed over several orders of magnitude, thus a log transformation is a good first step in the analysis. In addition, higher affinities produce lower K i values, thus it is valuable to calculate: pK i = 2log 10 (K i ). Higher affinities have higher pK i values, and each unit of pK i value corresponds to one order of magnitude of K i value. Table S4 presents the raw data transformed into pK i values. Generally, the highest K i value generated by NIMH-PDSP is 10,000, which produces a pK i value of 24 (although a value of 10,450 was reported for 5-MeO-TMT). For non-PDSP data gathered from the literature, some K i values greater than 10,000 are reported (i.e. 12,500, 14,142, 22,486, 39,409 and 70,000 for ibogaine).
When the primary assay did not produce .50% inhibition, the K i value is treated as .10,000. When the primary assay hit, but the secondary assay was not performed, the K i value is also treated as .10,000. If a secondary assay produced a K i value significantly greater than 10,000, it is usually also reported as .10,000. The lowest K i value in the data set of this study is 0.3 (lisuride at 5-HT 1A ) and the highest value is 70,000 (ibogaine at D 3 ), thus collectively, the data in this study cover nearly six orders of magnitude of K i values. However, ignoring values reported as .10,000, the K i values for a single drug in this study never exceed four orders of magnitude in range.
The goal of the normalization used in this study is to factor out potency, in order to allow easy comparison of the multi-receptor affinity profiles of different drugs. The normalization will adjust

average human
Receptor affinity data for morphine collected from the literature. The columns identify the receptor, the radioligand used in determining affinity, the source species from which the receptor was used, the tissue from which the receptor was used, the K i value in nanomoles or the IC50 (the molar concentration of an unlabeled agonist or antagonist that inhibits the binding of a radioligand by 50%, [26]) value in nanomoles, and the literature reference from which the data was obtained. doi:10.1371/journal.pone.0009019.t001 the highest pK i value for each drug to a value of 4, and set all K i values reported as .10,000 to a value of zero. K i values actually measured as greater than 10,000 are not set to zero (i.e. 5-MeO-TMT and ibogaine). We will call this normalized value npK i . Let the maximum pK i value for each drug be called pK iMax . For each individual drug: N If K i treated as .10,000, then npK i = 0 N npK i = 4+pK i 2pK iMax With this normalization: N higher affinities have higher values N affinities too low to be measured will be reported as zero N for each drug, the highest affinity will be set to a value of 4 N each unit of npK i value represents one order of magnitude of K i value N potency is factored out so that drugs of different potencies can be directly compared This normalization effectively factors out the absolute potency of each drug, and allows us to focus on the relative affinities of each drug at each receptor.

Perceptibility
It will also be seen that many psychedelic drugs interact with a large number of receptors. Fig. 3 shows the ranked distributions of npK i values for DOB and DOI, and the same data is listed below in numerical form (0.00 means K i .10,000, ND means the data is not available): DOB For potent compounds like DOB and DOI, it is possible to measure K i values over nearly a full four orders of magnitude range of affinity. However, not all of these affinities are able to produce perceptible mental effects. As a rule of thumb, 100-fold affinity is considered truly selective. Thus, receptors with npK i values below about 2.0 should not have perceptible mental effects. In Fig. 3, a black vertical bar represent a 100-fold drop in affinity relative to the receptor with the highest affinity, and divides those npK i values greater than 2.0 (on the left) from those 2.0 or less (on the right). This is presumed to be the limit of perceptible receptor interaction. Receptors to the right of the black bar should be imperceptible, while receptors to the left of the black bar should be perceptible, increasingly so the further left they are. In spite of the long tail of affinities, DOB is effectively selective for the three 5- Receptor affinity data for THC collected from the literature. The columns identify the receptor, the radioligand used in determining affinity, the source species from which the receptor was used, the tissue from which the receptor was used, the K i value in nanomoles, and the literature reference from which the data was obtained. doi:10.1371/journal.pone.0009019.t002 Figure 3. Receptor affinity profiles of DOB and DOI, ordered by decreasing affinity. The vertical axis is normalized pK i (npK i ). Horizontal axis is a list of forty-two receptors, arranged in order of decreasing affinity for each individual drug. Colors correspond to classes of receptors, and are the same as used in Fig. S1. The black vertical bars represent a 100-fold drop in affinity relative to the receptor with the highest affinity. As a rule of thumb, this is presumed to be the limit of perceptible receptor interaction. Receptors to the right of the black bar should be imperceptible, while receptors to the left of the black bar should be perceptible, increasingly so the further left they are. doi:10.1371/journal.pone.0009019.g003 HT 2 (serotonin-2) receptors (Beta2 falls at the approximate limit of perceptibility), while DOI by contrast has nineteen receptors in the presumed perceptible range, although they should not all be equally perceptible.

Breadth
An index of the breadth (or inverse of selectivity), B, of the binding profiles of the individual drugs or receptors can be constructed by summing the forty-two npK i values for each drug, or the thirty-five npK i values for each receptor. If a drug were absolutely selective, binding at only one receptor (e.g. salvinorin A), it would have the minimal B value of 4, regardless of the absolute affinity of the drug for its one receptor. If a drug bound with equal affinity to all forty-two receptors, it would have the maximum B value of 4642 = 168, regardless of its absolute receptor affinities.
It is not clear that a simple sum of npK i values is the best index of breadth. In this method, four receptors with K i values of 1,000 collectively carry the same weight as one receptor with a K i value of 1. This may not be a realistic equivalence. Thus we will include three measures of breadth: B sq and B exp give greater weight to higher affinity (lower K i ) values. Regression analysis of receptor affinity vs. potency in humans suggests that B sq is the most meaningful breadth statistic. Table S5 and Table S6 present the raw K i data converted into npK i values, for both the individual receptors, and groups of receptors summed using the B sq statistic.

Proportional Breadth
In addition to looking at the breadth of interaction of individual drugs with multiple receptors, it may be of value to look at an individual drug's interaction with one receptor or group of receptors, as a proportion of the drug's total interaction with all receptors.
In order to compute the proportion for and individual receptor or a group of receptors, we divide the sum of squares of npK i values for the group of receptors, by the sum of squares of npK i values for all receptors: For example, to compute this proportion for ''5-HT'' receptors, we divide the squares of the values in the ''5-HT'' column of Table S5 (for LSD, 11.13 2 = 123.9), by the squares of the values in the center column (''B sq '') of Table 3 (for LSD, 13.12 2 = 172.1); 123.9/172.1 = 0.719 for LSD. We will call this proportion B p . The proportional breadth data is displayed in Table S7 and  Table S8.

Truncated Receptor Profiles
The NIMH-PDSP generally does not measure K i values greater than 10,000 nm, because at those concentrations, there is a great deal of non-specific binding which invalidates the measurement of receptor affinity. This creates a problem for drugs whose best-hit has a K i value of greater that 100 nm (TMA, mescaline, TMA-2, DIPT, MDMA, 5-MeO-DIPT, ibogaine). For these drugs, the range of K i values that can be measured by the NIMH-PDSP is less than the 100-fold presumed perceptible range, and therefore, the lowest measurable npK i value is greater than the presumed limit of perceptibility at 2.00. Table 4 shows for each drug, the lowest K i value measured (K i Min) which is the best-hit, the best-hit receptor (K i MinR), the theoretically lowest measurable npK i value (npK i Lim), the lowest actually measured npK i value (npK i Min), and the receptor where the lowest npK i value was actually measured (npK i MinR). Drugs that have both a K i Min value of greater that 100 nm and an npK i Min value greater than 2.00 have truncated receptor affinity profiles. Seven drugs have best-hit K i values of greater that 100 nm: TMA, mescaline, TMA-2, DIPT, MDMA, 5-MeO-DIPT, ibogaine. For these seven drugs, the perceptible receptor profile is truncated due to the methodological limitations of the NIMH-PDSP, to the extent to which the npK i Min value is greater than 2.00. Note that for ibogaine, whose best hit is 206 nm, the npK i Min value is 1.47, indicating that the receptor profile is not fully truncated, because several K i values above 10,000 nm were gathered from the literature (however, ibogaine has not received a full receptorome screening, and thus its receptor profile must be considered incomplete for other reasons).
Six drugs have both a K i Min value of greater that 100 nm and an npK i Min value greater than 2.00. For 5-MeO-DIPT, the npK i Min value (2.15) is close to the presumed perceptibility limit, thus we can consider its receptor profile to be complete. For TMA-2, DIPT, and MDMA, the npK i Min values (2.58, 2.51, 2.43 respectively) fall in the weak region of the presumed perceptibility range. Although these three receptor profiles are truncated, the missing data may be of little consequence. For TMA and mescaline, the npK i Min values (2.98. 2.92 respectively) fall in the moderate region of the presumed perceptibility range. We must consider these two truncated receptor profiles to be truly incomplete, with potential consequences for our interpretations of the properties of these two drugs. The receptor profiles of some other drugs are incomplete due to holes in the NIMH-PDPS data set. Morphine and THC have not been broadly assayed, and must also be considered to be incomplete.

Normalized Affinity Data
Fig. S1 shows the simplest view of the normalized affinity data. The drugs in Fig. S1 are ordered to correspond roughly to similarity of structure and receptor affinity profiles. Colors correspond to classes of receptors. It can be seen at a glance that most, but not all of the drugs interact strongly with the serotonin receptors (beige), certain drugs interact strongly with the dopamine receptors (red), others with the adrenergic receptors (green), yet others with the histamine receptors (yellow), etc.

Breadth
In Table 3, the thirty-five drugs are arranged in order of decreasing breadth and increasing selectivity, based on the three breadth indices B, B sq , and B exp . Although the three indices provide different orderings, the orderings are quite similar at the two extremes of the table (greatest and least breadth) where most of the attention is likely to be focused. The drugs with the broadest receptor interactions (least selective) are found at the tops of the columns, and the drugs with the narrowest receptor interactions (most selective) are found at the bottoms of the columns. Regression analysis suggests that B sq is the best statistic for combining receptors, therefore the B sq statistic will be used in most of the breadth analyses to follow. The B, B sq and B exp data of Table 3 is presented graphically in Fig. 4.

Profiles of Drugs
The relative breadth or selectivity of the thirty-five drugs is nicely visualized in Fig. S2, in which for each drug, the bars representing forty-two receptors are arranged in order of decreasing size. The drugs are arranged in order of decreasing breadth, based on the B sq values of Table 3   It is also useful to present the npK i data of Fig. S2 in numerical format. In the listing below, the npK i values for each drug are arranged in decreasing order. A value of 0.00 means that the K i value was measured as .10,000 nm. ''ND'' indicates that the data is not available. The 5-HT 2A and 5-HT 2C receptors are also highlighted in bold font for easier location. npK i values below about 2.0 should be imperceptible, while values above about 2.0 should be perceptible, and the higher the npK i value, the more perceptible a receptor should be.

Groups of Related Receptors
In addition to looking at breadth at the full complement of fortytwo receptors with which the drugs interact, we can use the B sq statistic to look at the participation of selected groups of receptors (Table S5). 5-HT 6 and 5-HT 7 are grouped because they share a common G-protein, G s [13]. The same data is also presented in Table S6, which also includes interactions with individual receptors. Tables S5 and S6 allow us to easily identify drugs with the greatest or least interaction with an individual receptor or any group of related receptors.
For example, among drugs with measurable affinity at a 5-HT 2 receptor, 5-MeO-DMT has the weakest interaction with both the three 5-HT 2 receptors and the two paradigmatic 5-HT 2 receptors (5-HT 2A/C ). This can be seen in that 5-MeO-DMT has the lowest non-zero value in all four of the columns: ''5-HT 2 ,'' ''5-HT 2 max,'' ''5-HT 2A , 5-HT 2C ,'' and ''5-HT 2A/C max.'' Meanwhile, the drugs 2C-E and DOI rise to the top of the same four columns, indicating that they have the strongest interactions with the 5-HT 2 receptors (fourteen drugs tie for the top value in the 5-HT 2 max column because thirteen drugs have 5-HT 2B as their best hit and one has 5-HT 2C as its best hit). LSD has the strongest interaction collectively with the five dopamine receptors (D 1 , D 2 , D 3 , D 4 , D 5 ), the ten assayed 5-HT receptors, and the four assayed 5-HT 1 (serotonin-1) receptors. DMT has the strongest interaction with any single dopamine receptor (Dmax), and is the only drug to have its best hit at the 5-HT 7 receptor (Table S6). Mescaline is the only drug to have its best hit at an adrenergic receptor (a 2C , Table S6). Ibogaine is the only drug to have its best hit at a sigma receptor. MDMA is the only drug to have its best hit at an imidazoline receptor (Table S6). These observations likely provide clues to the qualitative diversity of these drugs.

Proportional Breadth
There is yet another way to look at the participation of the subsets of receptors. It may be relevant to consider the participation of a sub-set of receptors in proportion to the participation of all receptors (Table S7). For this we use the proportional breadth statistic (B p ) described in the methods section. The same data is also displayed in Table S8.
The proportional breadth statistic, B p , is strongly influenced by the degree of overall breadth of the drug (B sq 2 ), as this determines the denominator in calculating the proportion. Therefore, we find MEM at the top of the 5-HT column because it is almost completely selective for a single receptor, 5-HT 2B . Similarly, TMA-2 is at the top of the 5-HT 2 columns because it is highly selective for the 5-HT 2 receptors, and EMDT is at the top of the ''5-HT 6 , 5-HT 7 '' column because it is highly selective for the 5-HT 6 receptor. Due to their high degree of selectivity, MEM, TMA-2, and EMDT all have small denominators in calculating the B p statistic.
More interesting cases involve less selective drugs. For example, DMT is a very promiscuous drug, yet it falls at the top of the ''a 1A , a 1B '' column. MDA falls at the top of the ''Adrenergic'' and ''a 2A , a 2B , a 2C '' columns. 5-MeO-DMT appears third from the top of the ''5-HT 6 , 5-HT 7 '' column and second from the top of the ''5-HT 7 '' column (Table S8). Ibogaine appears at the top of the ''s 1 , s 2 '' column. Although fairly selective, TMA-2's position at the top of the ''H 1 , H 2 '' column represents an important aspect of its pharmacology, likewise for DOM's position at the top of the ''b 1 , b 2 '' column.

Profiles of Receptors
The B statistics can also be used to look at the relative role played by the various receptors in the pharmacology of the entire set of drugs of this study. In this case, for each receptor, we sum the npK i values at each receptor across each of the thirty-five drugs. In Table 5 we can see the rankings from the three B statistics. The data in Table 5 can also be represented graphically (Fig. 5). The three B statistics provide a very consistent ranking for the top seven receptors. In descending order of importance: 5-HT 2B , 5-HT 1A , 5-HT 7 , 5-HT 1D , 5-HT 2A , 5-HT 2C , a 2C (with some play between the sixth and seventh positions). This set of top receptors would be a good place to look for receptors other than 5-HT 2A and 5-HT 2C , which play an important role in the actions of psychedelic drugs. Fig. S3 is a more detailed graphical view of data presented in Table 5. The receptors are presented in order of decreasing breadth (B sq ). The figures for each individual receptor provide information on the relative importance of each receptor at each drug, similar to that in Table S5 and Table S6

Activity Data
The NIMH-PDSP provided activity data for the twenty-five drugs of Fig. 1, for 5-HT 2A and 5-HT 2C (Table S3). While most compounds appear to be full agonists at the two receptors, there are a few exceptions. Lower values of activity were reported for psilocin, MDMA, DOM and the three control drugs: 4C-T-2, 6fluoro-DMT, and lisuride.

Discussion
Perhaps the most striking result of the NIMH-PDSP assays has been to show that the psychedelics interact with a large number of receptors (forty-two of the forty-nine sites at which most of the drugs were assayed). While the phenylalkylamines tend to be more selective than the tryptamines and ergolines, they generally can not be accurately characterized as selective for 5-HT 2 , as they are so widely described in the literature [1,4,6,7,14,15]. Only DOB and MEM come close to fitting that description. Ironically, DOI has been one of the drugs of choice in studies of the molecular pharmacology of psychedelics, and has been widely assumed to be a 5-HT 2 -selective agent [4,6,7,15,16]. This study has revealed DOI to be one of the least selective of all psychedelics. Some of the literature on DOI may need to be reinterpreted. The same may be true of any studies whose conclusions rely on the assumption that psychedelics are selective. Studies requiring drugs selective for 5-HT 2 should be conducted with DOB or MEM, and they should not be presented as typical or characteristic of psychedelics.
In addition to showing that psychedelics are not as selective as generally believed, the data presented also shows that they exhibit diverse patterns of receptor interactions. Different drugs emphasize different classes of receptors. 5-HT 2B is the best hit for thirteen drugs, and 5-HT 1A is the best hit for nine drugs. Five of the top six psychedelic receptors (Table 5)  It should be possible to use this diverse set of drugs as probes into the roles played by the various receptor systems in the human mind. In the papers that follow, this possibility will be explored by synthesizing the NIMH-PDSP data together with the data on the human pharmacology of these drugs. Figure S1 Receptor affinity profiles of psychedelic drugs, ordered by receptor type. The vertical axis is normalized pKi (npKi). Horizontal axis is a list of forty-two receptors, grouped by receptor type. The drugs are ordered to correspond roughly to similarity of structure and receptor affinity profiles. Colors correspond to classes of receptors. It can be seen at a glance that most, but not all of the drugs interact strongly with the serotonin receptors (beige), certain drugs interact strongly with the dopamine receptors (red), others with the adrenergic receptors (green), yet others with the histamine receptors (yellow), etc. Found at: doi:10.1371/journal.pone.0009019.s001 (0.14 MB DOC) Figure S2 Receptor affinity profiles of psychedelic drugs, ordered by decreasing affinity. The vertical axis is normalized pKi (npKi). Horizontal axis is a list of forty-two receptors, arranged in order of decreasing affinity for each individual drug. The thirty-five drugs are arranged in order of decreasing breadth, based on the Bsq values of Table 3 and Fig. 4. Drugs at the top of the figure have the broadest receptor interactions (least selective), while drugs at the bottom of the figure have the narrowest receptor interactions (most selective). Colors correspond to classes of receptors, and are the same as used in Fig. S1. The black vertical bars represent a 100-fold drop in affinity relative to the receptor with the highest affinity. As a rule of thumb, this is presumed to be the limit of perceptible receptor interaction. Receptors to the right of the black bar should be imperceptible, while receptors to the left of the black bar should be perceptible, increasingly so the further left they are. Found at: doi:10.1371/journal.pone.0009019.s002 (0.18 MB DOC) Figure S3 Receptor affinities at forty-two receptors across thirtyfive drugs, ordered by decreasing breadth of receptor. The vertical axis is normalized pKi (npKi). Horizontal axis is a list of thirty-five drugs, ordered by decreasing affinity at the receptor. The forty-two receptors are arranged in order of decreasing breadth, based on the Bsq values of Table 5 Table S1 Receptor affinity data for ibogaine. Table S1 reports receptor affinity data for ibogaine collected from the literature. The columns identify the receptor, the species from which the receptor was used, the tissue from which the receptor was used, the radioligand used in determining affinity, the Ki value in nanomoles or the IC50 value in nanomoles, and the literature reference from which the data was obtained. Found at: doi:10.1371/journal.pone.0009019.s004 (0.31 MB DOC)

Table S2
Raw affinity (Ki) data for thirty-five drugs at sixty-seven sites. The table has been divided into three sections. The first section displays forty-two sites at which most compounds were assayed and at least one ''hit'' (Ki,10,000 nm) was found (5ht1a, 5ht1b, 5ht1d, 5ht1e, 5ht2a, 5ht2b, 5ht2c, 5ht5a, 5ht6, 5ht7, D1, D2, D3, D4, D5, Alpha1A, Alpha1B, Alpha2A, Alpha2B, Alpha2C, Beta1, Beta2, SERT, DAT, NET, Imidazoline1, Sigma1, Sigma2, DOR, KOR, MOR, M1, M2, M3, M4, M5, H1, H2, CB1, CB2, Ca+Channel, NMDA/MK801). The second section displays seven sites at which most compounds were assayed, but at which there were no hits (5ht3, H3, H4, V1, V2, V3, GabaA). The third section displays the remaining eighteen sites, at which only a few compounds were assayed, and no hits were found (GabaB, mGluR1a, mGluR2, mGluR4, mGluR5, mGluR6, mGluR8, A2B2, A2B4, A3B2, A3B4, A4B2, A4B2**, A4B4, BZP (a1), EP3, MDR 1, PCP). Missing Ki values are indicated by ''ND'' meaning that no data is available, or by ''PH'' meaning that the primary assay ''hit'' (.50% inhibition), but the secondary assay was not performed. For the first section of the table, an extra row and column labeled ''ND/PH'' provides a count of missing Ki data in the two categories, for each compound and for each receptor.  Table S3 Activity data for twenty-five drugs at 5-HT2A and 5-HT2C. GF62 is the cell line that expresses the 5-HT2A receptor, and INI is the cell line that expresses the 5-HT2C receptor. The ''EC50 nM'' columns express the concentration that gives half of the maximal activity for that drug. The maximal activity is displayed in the ''Emax6SEM'' column, and represent Ca++ mobilization relative to 5-HT which should give an Emax value of 100%. Data for the drugs should produce lower Emax values. For a compound that gives, for example, 53% Emax, the EC50 is the concentration where 26.5% response occurs. Emax values above 100%6SEM are an artifact caused by extrapolation by the graphpad program when it doesn't have points at the top end to define the asymptote. The data represent the mean 6 variance of computer-derived estimates from single experiments done in quadruplicate. Thus, the four observations are averaged and a single estimate with error is provided.   Table S5 Thirty-five drugs arranged in order of decreasing breadth at selected groups of receptors. The thirty-five drugs are arranged in order of decreasing breadth at groups of receptors, based on the breadth index Bsq. The drugs with the broadest receptor interactions within the group are found at the tops of the columns, and the drugs with the least receptor interactions are found at the bottoms of the columns. Some columns list the maximum npKi value for a group of receptors (e.g. 5-HT2max). In this example, the column lists the highest npKi value of the three 5-HT2 receptors (5-HT2A, 5-HT2B, and 5-HT2C). An entry of ''ND'' indicates that the statistic could not be calculated because some data is missing. For example, to calculate Bsq for 5-HT2, or npKi for 5-HT2max, we need npKi values for 5-HT2A, 5-HT2B, and 5-HT2C. If any one of the three values is missing, the statistics will be reported as ND. Values of 0.00 correspond to Ki values of .10,000. The same data is also presented in Table S6 Receptors in a group are listed in the column heading, or are represented with the following abbreviations: N 5-HT -5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E, 5-HT2A, 5-HT2B, 5-HT2C, 5-HT5A, 5-HT6, 5-HT7 N 5-HT2 -5-HT2A, 5-HT2B, 5-HT2C N 5-HT2max -maximum of 5-HT2A, 5-HT2B, 5-HT2C N 5-HT2A/Cmax -maximum of 5-HT2A, 5-HT2C N 5-HT1 -5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E N 5-HT1max -maximum of 5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E N Dmax -maximum of D1, D2, D3, D4, D5 N Adrenergic -a1A, a1B, a2A, a2B, a2C, b1, b2 N AdrenergicMax -maximum of a1A, a1B, a2A, a2B, a2C, b1, b2 N a1max -maximum of a1A, a1B N a2max -maximum of a2A, a2B, a2C N bmax -maximum of b1, b2 N Hmaxmaximum of H1, H2 N smax -maximum of s1, s 2 N Mmaxmaximum of M1, M2, M3, M4, M5 N TransportersMaxmaximum of SERT, DAT, NET N OpioidMax -maximum of DOR, KOR, MOR Found at: doi:10.1371/journal.pone.0009019.s008 (0.61 MB DOC)  Table S6 presents the normalized pKi data (npKi) for thirty five drugs at forty-two receptors, transporters and ion channels. In addition, it presents the three breadth statistics (B, Bsq, Bexp) for each drug across all forty-two sites, the breadth statistic Bsq for sixteen groups of related sites, and the maximum npKi value for thirteen groups of related sites. A useful way to work with the table is to choose a column, and sort the data on that column (click ''Data, ''Sort,''  Table S7 Thirty-five drugs arranged in order of decreasing proportional interaction at selected groups of receptors. The thirty-five drugs are arranged in order of decreasing proportional interaction at groups of receptors, based on the proportional breadth index Bp. The drugs with the greatest proportional interactions are found at the tops of the columns, and the drugs with the least proportional interactions are found at the bottoms of the columns. An entry of ''ND'' indicates that the statistic could not be calculated because some data is missing. The same data is also presented in Table S8. Receptors in a group are listed in the column heading, or are represented with the following abbreviations: N 5-HT -5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E, 5-HT2A, 5-HT2B, 5-HT2C, 5-HT5A, 5-HT6, 5-HT7 N 5-HT1 -5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E N Adrenergic -a1A, a1B, a2A, a2B, a2C, b1, b2 Found at: doi:10.1371/journal.pone.0009019.s010 (0.37 MB DOC) Table S8 Table of Bp values at individual or groups of receptors  for thirty-five drugs. Table S8 presents the proportional breadth statistic Bp for each drug at each of forty-two sites, and for sixteen groups of related sites. A useful way to work with the table is to choose a column, and sort the data on that column (click ''Data, ''Sort,'' select ''Header row,'' choose the column from the ''Sort by'' drop down and select ''Descending,'' finally clicking ''OK''). Missing values are reported as ''ND.'' Receptors in a group are listed in the column headings using the following abbreviations: