23Drug-resistant mutations often have deleterious impacts on replication fit-24 ness, posing a fitness cost that can only be overcome by compensatory muta-25 tions. However, the role of fitness cost in the evolution of drug resistance has 26 often been overlooked in clinical studies or in vitro selection experiments, as 27 these observations only capture the outcome of drug selection. In this study, 28 we systematically profile the fitness landscape of resistance-associated sites in 29 HIV-1 protease using deep mutational scanning. We construct a mutant li-30 brary covering combinations of mutations at 11 sites in HIV-1 protease, all of 31 which are associated with resistance to protease inhibitors in clinic. Using deep 32 sequencing, we quantify the fitness of thousands of HIV-1 protease mutants 33 1 after multiple cycles of replication in human T cells. Although the majority 34 of resistance-associated mutations have deleterious effects on viral replication, 35 we find that epistasis among resistance-associated mutations is predominantly 36 positive. Furthermore, our fitness data are consistent with genetic interactions 37 inferred directly from HIV sequence data of patients. Fitness valleys formed 38 by strong positive epistasis reduce the likelihood of reversal of drug resistance 39 mutations. Overall, our results support the view that strong compensatory 40 effects are involved in the emergence of clinically observed resistance muta-41 tions and provide insights to understanding fitness barriers in the evolution 42 and reversion of drug resistance. 43 1 Introduction 44 Antibiotics and antiviral drugs have achieved great success in recent history [1].
45However, therapeutic failure may occur due to low adherence and the emergence of 46 drug resistance [2, 3]. The increasing amount of drug resistant pathogens is a global 47 threat to public health [4][5][6][7][8][9][10][11]. The genetic barrier to drug resistance, defined as 48 the number of mutations needed to acquire resistance, is a major determining factor 49 of treatment outcomes [12][13][14]. Another important but often overlooked aspect of 50 drug resistance is the fitness barrier [15][16][17]. Drug resistance associated mutations 51 (DRAMs) in pathogen proteins may decrease enzymatic activities, interfere with 52 molecular interactions, or destabilize the protein structure [18][19][20][21][22]. Because of the 53 impaired replication capacity without drug selection, drug-resistant mutants can-54 not normally outcompete wild-type or establish in the population [23][24][25]. However, 55 drug-resistant mutants can sometimes reach substantial frequency in the population. 56 Fluctuating drug concentrations may create time windows when drug-resistant mu-57 tants replicate better than wild-type virus [26]. Moreover, compensatory mutations 58 can rescue the impaired replication capacity of mutants and stabilize drug resis-59 tance [22, 27, 27-29]. Thus, comprehensive quantification of the fitness landscape is 60 needed to predict the evolution of drug ...