Inhibition of the Ethanol-induced Potentiation of α1 Glycine Receptor by a Small Peptide That Interferes with Gβγ Binding*

Background: Gβγ interaction with GlyR is an important determinant in ethanol potentiation of this channel. Results: A small peptide, RQHC7, can inhibit ethanol potentiation of GlyR currents. Conclusion: Results with RQHC7 indicate that ethanol mediated potentiation of GlyR is in part by Gβγ activation. Significance: Molecular interaction between Gβγ and GlyR could be used as a target for pharmacological modification of ethanol effects. Previous studies indicate that ethanol can modulate glycine receptors (GlyR), in part, through Gβγ interaction with basic residues in the intracellular loop. In this study, we show that a seven-amino acid peptide (RQHC7), which has the primary structure of a motif in the large intracellular loop of GlyR (GlyR-IL), was able to inhibit the ethanol-elicited potentiation of this channel from 47 ± 2 to 16 ± 4%, without interfering with the effect of Gβγ on GIRK (G protein activated inwardly rectifying potassium channel) activation. RQHC7 displayed a concentration-dependent effect on ethanol action in evoked and synaptic currents. A fragment of GlyR-IL without the basic amino acids did not interact with Gβγ or inhibit ethanol potentiation of GlyR. In silico analysis using docking and molecular dynamics allowed to identify a region of ∼350Å2 involving aspartic acids 186, 228, and 246 in Gβγ where we propose that RQHC7 binds and exerts its blocking action on the effect of ethanol in GlyR.

Previous studies indicate that ethanol can modulate glycine receptors (GlyR), in part, through G␤␥ interaction with basic residues in the intracellular loop. In this study, we show that a seven-amino acid peptide (RQH C7 ), which has the primary structure of a motif in the large intracellular loop of GlyR (GlyR-IL), was able to inhibit the ethanol-elicited potentiation of this channel from 47 ؎ 2 to 16 ؎ 4%, without interfering with the effect of G␤␥ on GIRK (G protein activated inwardly rectifying potassium channel) activation. RQH C7 displayed a concentration-dependent effect on ethanol action in evoked and synaptic currents. A fragment of GlyR-IL without the basic amino acids did not interact with G␤␥ or inhibit ethanol potentiation of GlyR. In silico analysis using docking and molecular dynamics allowed to identify a region of ϳ350Å 2 involving aspartic acids 186, 228, and 246 in G␤␥ where we propose that RQH C7 binds and exerts its blocking action on the effect of ethanol in GlyR.
The glycine receptor (GlyR) 2 is a ligand-gated ion channel (LGIC) that has critical inhibitory functions in the spinal cord, brain stem, and some upper brain regions. Its activation by the neurotransmitter glycine increases its permeability to chloride ions (Cl Ϫ ) (1), producing a fast inhibition of action potential generation in the postsynaptic membrane. GlyRs are involved in several important physiological processes such as motor and respiratory rhythm regulation and pain perception respiratory rhythm regulation and pain perception (1), being ethanol, due to its biomedical relevance, one of the most important glycinergic modulators. To its biomedical relevance, ethanol is one of the most important glycinergic modulators. In addition, it is well accepted that this drug of abuse has a potent action on inhibitory channels, such as GlyR and ␥-aminobutyric acid receptors, where it induces a robust potentiation of their channel activities. Along with the N-methyl D-aspartate (NMDA) and 5-hydroxytryptamine type 3 receptors (2), these LGICs offer the most current molecular explanation for the well known depressing effects that this drug induces on animal behavior.
As a common feature of a member of the Cys-loop family of LGIC, GlyR have a quaternary structure composed of five subunits arranged around a central pore. Four ␣ (␣1-4) and only one ␤ subunit have been described. Each subunit possesses an extracellular N-terminal domain, four transmembrane domains (TM 1-4) and a large intracellular loop (IL) between TM 3 and 4 (1). Although this cytoplasmic domain is a common feature in the LGIC family, its primary sequence is poorly conserved. In addition, the structural and functional properties of this intracellular domain are not well understood. Nevertheless, it was reported that it is important for ion permeation in the Torpedo acetylcholine receptor (3). In addition, the GlyR-IL has been found to have a role in G protein modulation (4 -6), sorting (7,8), synaptic clustering (9), and receptor phosphorylation (10 -12).
G proteins are important intracellular regulators of many signal transduction pathways, and their function is regulated by the guanine nucleotide exchange activity of activated receptors and other proteins, such as RGS (regulators of G protein signaling), GDI (GDP dissociation inhibitor) and GAP (GTPase-activating protein). Therefore, in the activated state, both G␣ and G␤␥ subunits are able to modulate multiple effector proteins (13). G protein-regulated pathways are recognized as one of the most important in terms of hormonal cell signaling mechanisms (14), and the pharmacological modification of these pathways has been a central focus of research. In this context, the use of small peptides to induce molecular interference of protein-protein interactions has been of importance in the study of molecular events in signal transduction pathways (15). However, not much is known regarding a potential pharmaco-logical modulation of ion channels and ethanol modulation using small peptides (20). Furthermore, this strategy has been applied to the inhibition of G␤␥-mediated activation of adenylyl cyclase, GIRK channels (16), and phospholipase C (17)(18)(19).
In this report, we show that a seven-amino acid peptide (termed RQH C7 ) is capable of interfering with G␤␥ binding to the GlyR-IL and inhibiting ethanol potentiation. In terms of selectivity, the peptide did not have effects on another G␤␥ effector, the activation of GIRK through a GABA B agonist. Moreover, the potentiation of the synaptic activity induced by ethanol was inhibited in the presence of RQH C7 . Finally, the use of in silico techniques predict that this peptide binds with high affinity to a site in G␤␥ where this protein would interact with GlyR.
GST Pulldown Assays-Fusion protein expression and GST pulldown assays were performed as described previously (20). Briefly, DNA fragments encoding the cytoplasmic domain of GlyR were subcloned in the vector pGEX-5X3 (GE Healthcare). GST fusion proteins were expressed in Escherichia coli BL21 bacteria using 50 M isopropyl 1-thio-␤-D-galactopyranoside. Subsequently, the proteins were purified using a glutathione resin (Novagen). Normalized amounts of GST fusion protein were incubated with purified G␤␥ protein (10 ng,Calbiochem) in the presence and absence of 2 M of peptides RQH, RQH C10 , RQH C7 , and RQH N . After several washing steps, bound proteins were separated on 10% SDS-polyacrylamide gels, and G␤␥ binding to GlyR-IL was detected using an anti-G␤ antibody (1:1000, Santa Cruz Biotechnology) and a chemiluminescence kit (PerkinElmer Life Sciences). Finally, the relative amount of G␤␥ was quantified by densitometry. G␤ detection in Western blots was defined as a measure of GlyR-IL-G␤␥ binding.
Electrophysiology-For experiments with GIRK channels, HEK 293 cells were cultured using standard methodologies and co-transfected with plasmids encoding the GABA B receptor subunits GABA B1 (fused to GFP), GABA B2 , GIRK 1 , and GIRK 4 , using an Xfect transfection reagent kit (Clontech). Expression of GFP was used as a marker for positively transfected cells, and recordings were made after 18 -24 h. For glycine-evoked currents and synaptic activity, primary cultures of spinal cord neurons were obtained from 13-14 days mouse embryos (strain C57BL/J6). Whole-cell recordings were performed using a holding potential of Ϫ60 mV. Patch electrodes were filled with the following: 140 mM KCl, 10 mM BAPTA, 10 mM HEPES (pH 7.4), 4 mM MgCl 2 , 2 mM ATP, and 0.5 mM GTP, with or without 200 M RQH peptides. The external solution contained the following: 150 mM NaCl, 5.4 mM KCl, 2.0 mM CaCl 2 , 1.0 mM MgCl 2 , 10 mM HEPES (pH 7.4), and 10 mM glucose. In the case of GIRK channel activity, KCl concentration was 120 and 30 mM in the internal and external solutions, respectively. Baclofen (10 M) was applied in short pulses (4 -5 s) every 2 min during 8 min. For the recording of ethanol-mediated potentiation of GlyR, a previously described methodology was used (6,21). Ethanol (100 mM) was co-applied with glycine (15 M), and the results were expressed as percentage of potentiation at 15 min. Although potentiation can be detected with 10 mM ethanol (25), we utilized 100 mM in most experiments to facilitate the statistical analysis. Synaptic activity was recorded after 13-16 days of in vitro culture. 6-cyano-7-nitroquinoxaline-2,3dione (4 M), bicuculline (4 M), and tetrodotoxin (100 nM) were applied to the external solution to isolate miniature glycinergic events.
Docking and Molecular Dynamics-An ␣ helix representation of the peptide was created based on the phi and psi angles calculated for crystallized proteins stored in the Protein Data Bank. The structure was minimized using molecular mechanics by the steepest descent method on CHARMM force field implemented in NAMD software. Docking and clustering were performed as described in Guzman et al. (20). Briefly, docking was performed with Zdock to obtain 2000 complex structures, which were grouped in 20 clusters by ClusPro (version 1.0). The structure for G␤␥ was obtained from Protein Data Bank (1TGB). For molecular dynamics, the representative complex of the most populated cluster was solvated in a cubic water box (90.3 ϫ 75.4 ϫ 70.7 Å) using water model TIP3P. These calculations were performed in NAMD (22) considering a CHARMM force field, 20-ns total time, integration time of 2 fs, 12 Å cut-off, and 10000 minimization previous steps using the steepest descendent method. The final structure of molecular dynamics was taken to perform a free energy calculation using FastContact and distance measures by VMD software (23). Root mean square deviation was computed first by aligning only the interacting residue with the average positions for each of the atoms involved (data not shown). 3000 frames were considered for this calculation. Then, deviation was calculated from distances based on changes of side chain atom positions through time by VMD software (24).
Data Analysis-Statistical analyses were performed using analysis of variance, and the results are expressed as the arithmetic mean Ϯ S.E. Values of p Ͻ 0.05 were considered statistically significant. Origin (MicroCal, version 6.0) software was used for all statistical analyses.

RESULTS
Regulation by signaling pathways has been described for several members of LGICs, including the effects of ethanol and G proteins on GlyRs (6,21,25). In the case of GlyR, it has been previously described that the GlyR-IL contains motifs that are important for G protein-and ethanol-mediated potentiation. For example, it was found that a stretch of residues located in the N-terminal region of this domain is important for the interaction with G␤␥ (20). In addition, electrophysiological studies demonstrated that both GTP␥S and ethanol potentiation of GlyR were affected by the mutation of residues 316 -320 pres-ent in this domain (21). The importance of these residues was also shown by direct in vitro interaction between the GlyR-IL and G␤␥ using GST pulldown experiments (20).
In the present study, we used the region from amino acids 309 to 325 in the GlyR-IL to generate a series of peptides (Fig. 1). Subsequently, to study the capacity of these peptides at inhibiting the binding between G␤␥ and GlyR-IL, we used an in vitro assay that utilizes the GST-GlyR-IL fusion protein, G␤␥ and the small peptides. The previously studied RQH peptide (that has the entire 309 -325 region) was included as a positive control for inhibition of the interaction between G␤␥ and GST-IL (20). The data showed that both RQH C10 and RQH C7 diminished the binding of G␤␥ to the GlyR-IL to 28 Ϯ 18 and 36 Ϯ 21% of control, respectively. However, the data show that RQH N did not affect the interaction of G␤␥ with GST-IL (Fig. 2). These results demonstrate that there is a region in the IL that interacts with G␤␥ under in vitro conditions and correlates well with previous functional studies (21).
To obtain information about the region of G␤␥ that might be involved in the interaction with RQH C7 , in silico studies were performed. Previous modeling assays determined that the RQH C7 structure is most likely ␣ helix (20). In addition, docking, clustering, and molecular dynamics were done to determine the interaction nature of RQH C7 and G␤␥. The results of these analyses identified a region of ϳ350 Å 2 in G␤␥ characterized by the presence of three acidic residues, Asp-186, -228, and -246, which might serve as the binding region to the peptide, and the position that RQH C7 adopted on the G␤␥ molecule agrees with data obtained for the longer RQH peptide (Fig. 3, A  and B) (20). In addition, the data show that these residues can potentially form salt bridges with Arg-3, -6, and -7 in RQH C7 (20) due to the close distance between the residues (from 2.5 to 3.2 Å, Fig. 3C). Free energy calculations and atomic distances between these interacting residues further supported the presence of a binding region. To have an idea about deviations in atom positions of interacting amino acids, Arg-3, -6, -7 (in RQH C7 ) and Asp-186, -228, and -246 (from G␤␥) were considered for further molecular dynamic studies. The results were analyzed using a root mean square deviation (Fig. 3D), which showed that deviations in the position of side chains were Ͻ2 Å, which supports the notion that electrostatic interactions maintain the shape of the interface between RQH C7 and G␤␥.
It was recently shown that ethanol potentiation of GlyR is affected by G␤␥, via interaction with the GlyR-IL (21). Because RQH C7 was able to inhibit the binding of G␤␥ to the GlyR-IL, we decided to study whether this peptide was able to inhibit the ethanol potentiation of the GlyR. Patch clamp assays were performed in cultured spinal cord neurons expressing the ␣1␤ subunits (26), and the small peptides (at 200 M) were included into the recording pipette to allow them to diffuse into the cytoplasm. Glycine (15 M, at EC 15 ) and ethanol (100 mM) were extracellularly perfused to the cells to activate GlyR. After 15 min of internal solution dialysis, glycinergic responses were recorded in the absence and presence of ethanol (Fig. 4). In the control condition, ethanol potentiated the GlyR current to 47 Ϯ 2% above control, in agreement with previous reports (6,20,21). However, use of RQH, RQH C10 , and RQH C7 peptides reduced the potentiation to 19 Ϯ 4%, 17 Ϯ 1%, and 16 Ϯ 4% of control, respectively. On the other hand, RQH N , a peptide that lacks the basic motif (Fig. 1), did not inhibit this effect showing that RQH C7 is the shorter peptide containing the motif to inter-

FIGURE 2. Inhibition of the G␤␥-GlyR-IL binding by small peptides. A,
Western blot analysis of GST pulldown assays where binding of G␤␥ to GlyR-IL was examined. The G␤ immunoblot signal (upper panel) was considered a measure of G␤␥ binding to GlyR-IL (lower panel). The peptides (2 M) were included in the assay to test their capacity to interfere with this binding. Separated assays were performed to test each peptide. The binding of G␤␥ to GlyR-IL in absence of any peptide was taken as control. B, quantification of the G␤ binding to GlyR-IL. The G␤ signal obtained from the immunoblot for G␤ protein was normalized against the GST fusion protein present in the assay. An asterisk denotes statistical difference, p Ͻ 0.05 (n ϭ 4).
act with G␤␥ to inhibit ethanol potentiation (Fig. 4). Furthermore, the RQH C7 peptide inhibited the ethanol potentiation in a concentration-dependent manner where 20, 100, and 200 M of the peptide in the internal solution reduced the ethanol potentiation to 45 Ϯ 5, 29 Ϯ 1, and 16 Ϯ 4% above control, respectively (Fig. 5B). In addition, the inhibitory effect of RQH C7 was also detected using a lower ethanol concentration of 50 mM, which potentiated the current to 29 Ϯ 4% above control versus 13 Ϯ 1% with intracellular diffusion of RQH C7 (Fig. 5C).
Previous results from our laboratory showed that RQH also interfered with the signaling of G␤␥ in GIRK channels after stimulation of the GABA B receptor (20). In the present study, we reduced that peptide to a core of seven amino acids to generate a more specific effect on GlyR in comparison with GIRK. This possibility was studied in HEK cells that were transfected with both GABA B receptors and GIRK channels (Fig. 6). In this experiment, a specific agonist, baclofen, activates GABA B receptors leading to activation of a G i protein. The G␤␥ dimer of this protein opens the GIRK channel. In this experiment, baclofen was applied by external perfusion and 200 M RQH, or RQH C7 , were applied in the internal solution during 8 min. The data show that the potassium current was inhibited by RQH to 41 Ϯ 1% of the initial current. On the other hand, RQH C7 was unable to inhibit the GIRK current suggesting that this sevenamino acid motif induced a more specific inhibition of G␤␥ modulation to GlyR in comparison with RQH.
The above results were performed inducing an evoked Cl Ϫ current by application of exogenous glycine. To establish if RQH C7 had effects on ethanol alteration of synaptic parameters, we recorded glycinergic inhibitory post-synaptic currents (mIPSC) pharmacologically isolated in cultured spinal cord neurons. Among the parameters of the synaptic currents that were analyzed, the decay time constant was associated to the time that the single channels were maintained opened after synaptic stimulation. The analysis of this glycinergic activity showed that the decay time constant of the synaptic currents was increased by ethanol in a 42 Ϯ 9% (from 12 Ϯ 0.9 to 17 Ϯ 1.6 FIGURE 3. Region of molecular interaction existing between G␤␥ and RQH C7 . The equilibrated structure from 20 ns molecular dynamics shown as a zoom-out of the position where RQH C7 interacts with G␤␥ (A), and zoom-in of interacting residues of the peptide; Arg-3, -6, and -7 from RQH C7 and Asp-186 and -228 of G␤␥, respectively (B). Details (dashed rectangles in B) of the calculated interacting distances (in Armstrong) corresponding to saline bridges between the mentioned amino acids (C). Root mean square deviation (r.m.s.d.) was calculated considering distances from an average structure of side chains having 3000 frames representative of the 20-ns molecular dynamics (D).  (Fig. 7). Interestingly, in presence of the RQH C7 peptide, ethanol was unable to produce this enhancement and induced an increase of only 9 Ϯ 0.9% above control, which was not different from the control condition without ethanol. Furthermore, in presence of RQH N , which is devoid of basic residues, ethanol induced an increment in the decay time constant of 30 Ϯ 8%. Thus, these results show that RQH C7 inhibits the effects of ethanol on the glycinergic transmission.

DISCUSSION
Despite the efforts to ameliorate the impact of its use, ethanol consumption continues to be a major biomedical and social problem. In large part, this is because there are not very efficacious pharmacological means to interfere with its effects. One of the best understood ethanol effects is the potentiation of inhibitory GlyR, either by altering transmembrane domains or affecting the G protein-mediated regulation of this receptor (25). More recent studies have indicated that basic amino acids in the cytoplasmic domain of the GlyR are important for the latter mechanism (21).
Direct interaction between the GlyR-IL and G␤␥ was demonstrated in GST pulldown studies and is probably responsible for the electrophysiological effects of ethanol (20). For example, it was shown that the RQH peptide interfered with G␤␥ binding to the GlyR-IL (20). In the present study, we reported that similar inhibition was produced by two smaller peptides, RQH C10 and RQH C7 . Interestingly, RQH N , the amino-terminal sequence of RQH, did not cause the same interference indicating that the basic amino acids of Arg and Lys in the carboxyl terminus of the fragment are the critical residues for this interaction. Thus, residues from Arg-309 to Leu-315 are not critical for the interaction of G␤␥ with this segment of the GlyR-IL, which is in agreement with previous mutational studies on these residues (21,27). We demonstrated that RQH C7 (residues 316 to 322 of GlyR-IL) represents the smaller sequence from the peptides used able to bind G␤␥ causing an inhibition with GlyR binding. In agreement with this idea, in silico studies suggest that binding of the RQH C7 peptide was in a region that presented several acidic residues, which agrees with previous studies using RQH (20). Interestingly, free energy calculation obtained for RQH C7 was very similar to the one computed for SIGK (Protein Data Bank code 1XHM), another previously described G␤␥ interacting peptide (18) (Ϫ24 kcal/mol and Ϫ28 kcal/mol, respectively). Furthermore, the identified region of G␤␥ is functionally relevant because it contains motifs that interact with other effectors, namely GIRK, adenylyl cyclase, phospholipase C, and the ␣ subunit of G proteins (28).
In previous reports, it has been shown that the ␣1 GlyR subunit is potentiated by ethanol at a concentration of 10 mM and above (21,25). In this report, we used 50 and 100 mM of ethanol inducing potentiations of 29 and 45% above control, respectively (Fig. 5). Furthermore, it has been previously demonstrated that ethanol potentiation was inhibited by the use of G␤␥ scavengers (21). The peptides RQH, RQH C10 , and RQH C7 derived from the IL sequence also interfered with the ethanol effect on GlyR, suggesting that the molecular mechanism likely involves its binding and blocking of the G␤␥ region that interacts with the GlyR-IL. Interestingly, inhibition of the ethanol effect by RQH C7 was also elicited at a pharmacologically relevant concentration (50 mM), creating the possibility of a potential target for ethanol toxicity.
G␤␥ is a critical signaling modulator that has many important targets such as , phospholipase C, calcium, and potassium channels (29). It is recognized that the binding region of these G␤␥ targets share a common region in one interface of G␤␥, and this close molecular proximity might hinder the design of a specific ligand that interfere in the G␤␥-GlyR protein-protein interaction. For instance, a previous study demonstrated that the 17-amino acid RQH peptide, besides blocking the effect of ethanol on GlyR, also interfered with the G␤␥-dependent activation of GIRK channels (20). In the present study, supporting the idea that molecular selectivity can be achieved, we found that the shorter peptide RQH C7 did not interfere with this effect of G␤␥ on GIRK channels, confirming that the seven-amino acid peptide (RQH C7 ) displayed an improved specificity.
It is likely that, in part, the effects of ethanol in motor control are mediated by the potentiation of glycinergic transmission in spinal cord and brainstem synapses (30,31). In our hands, ethanol can affect a population of spinal neuron synapses containing mature ␣1 GlyRs. We analyzed the parameters of frequency, amplitude, rise time and decay time constant of the mIPSCs. The frequency of the synaptic events reflects a presynaptic measure of vesicular release. On the other hand, it is believed that the other parameters depend on the properties of the postsynaptic receptors such as the number and kinetic properties (32,33). We showed that ethanol induces an increase in decay time constant consistent with a postsynaptic action, and this effect was reverted almost completely by the intracellular perfusion of the RQH C7 peptide. We postulate that these data represent the intracellular interference of G␤␥ signaling that ultimately causes the inhibition of the ethanol synaptic potentiation of the glycinergic currents.
Peptides have been used as functional probes to study GPCR signaling pathways to elucidate their molecular dynamics, structural-functional organization, complex formation, and their role in the control of signaling. This was the case for peptides derived from the intracellular domain of dopamine (D 1 -DR, D 2 -DR), serotonin (5-HT 6 R, 5-HT 1B R), and acetylcholine (m 4 -MChR, m3-MChR) receptors, that selectively bind and are able to activate their respective G proteins, resembling the activity of the activated receptor (35)(36)(37)(38). Furthermore, pep- Glycine was used at a concentration equivalent to an EC 15 (15 M) and ethanol at 100 mM and 50 mM as indicated. A, evoked Cl Ϫ currents were recorded in the absence and presence of ethanol (100 mM) and RQH C7 at three concentrations. B, the percentage of ethanol potentiation was calculated with respect to the current obtained in the absence of ethanol. C, effects of RQH C7 on potentiation produced by 50 mM ethanol. RQH C7 peptide (200 M) was applied intracellularly through the recording pipette into spinal cord neurons. The percentage of ethanol potentiation was calculated with respect to the current obtained in the absence of ethanol and peptides. Analysis of variance and t-student statistical methods were applied. ***, p Ͻ 0.005; **, p Ͻ 0.01. tides derived from ␣2 and ␤2 adrenergic receptors stimulate G i/o and G s , respectively, and are believed to have potential pharmacological actions (37, 39 -41).
G␤␥ signaling is recognized to be important in several physiopathological conditions such as heart failure (42), tumorigenesis (43), cell migration and invasion (44), pharmacological tolerance and dependence (45), inflammatory processes (46), and several endocrine pathologies (47,48), among others. One of the first attempts at blocking the function of G␤␥ with selective peptides was developed for adenylyl cyclase. Chen and co-workers (16) inhibited the G␤␥ stimulated activity of a type 2, phospholipase C, ␤-adrenergic receptor kinase, and GIRK with a small peptide (QEHA) derived from the C2 domain of this adenylyl cyclase. Similar results were obtained with peptides derived from GIRK (50). Furthermore, Smrcka and coworkers (17,18,34,51,52) used a phage peptide display technique for screening peptides that bind to G␤␥. From these studies, the existence of a G␤␥ scavenger that can have a potential use in heart failure and be a synergistic aid to ␤-adrenergic receptor blockers is now recognized (49). Interestingly, the concentrations of RQH C7 (20 to 200 M) used in this report are comparable with the ones used for the QEHA peptide (300 M) (16) and is close to the same order of magnitude as SIRK (5 to 10 M) (51). Future studies will aim to identify small molecule to  circumvent the difficulties of working with peptides as therapeutic molecules.
Finally, we propose that the use of G␤␥-blocking peptides to interfere with pharmacologically altered glycinergic synapses might prove to be a novel approach to modify alcohol consumption and its associated health problems.